53
What does an LBO signal for the target’s industry? * Jarrad Harford Foster School of Business University of Washington [email protected] Jared Stanfield UNSW Business School UNSW Australia [email protected] Feng Zhang David Eccles School of Business University of Utah [email protected] This Draft: August 2015 Abstract This study examines the information that an LBO signals for the target’s industry. We show that LBOs tend to lead merger waves, and in particular predict more LBOs and strategic M&As in the same industry. We establish the valuation implications of this predictability for the LBO target’s industry as a whole. While the industry becomes more valuable following LBOs, surprisingly this is not fully impounded into stock prices at LBO announcement. Instead, as further merger activity occurs, the prices rise, leading to significant abnormal long run returns for peers of an LBO target. We attempt to distinguish between the hypothesis that LBO sponsors simply select into undervalued industries and/or industries that are changing as opposed to causing the changes. We find that our evidence is most consistent with causation, albeit there is some limited evidence of selection. Our study sheds light on the role of LBOs in merger waves as well as the motivation and value-drivers of LBOs. Keywords: leveraged buyouts, mergers and acquisitions, merger waves JEL classification: G14; G34 * We thank Hank Bessembinder, Audra Boone, Karl Lins, Ron Masulis, Konark Saxena, Robert Tumarkin, and Jason Zein, as well as the seminar participants at Australian National University, Deakin University, Fudan University, Massey University, Nanyang Technological University, the University of South Australia, University of Utah, UNSW Australia, and Utah State University for their useful comments. All remaining errors are our own.

What does an LBO signal for the target’s industry? · PDF fileWhat does an LBO signal for the target’s industry? Abstract This study examines the information that an LBO signals

Embed Size (px)

Citation preview

Page 1: What does an LBO signal for the target’s industry? · PDF fileWhat does an LBO signal for the target’s industry? Abstract This study examines the information that an LBO signals

What does an LBO signal for the target’s industry?*

Jarrad Harford Foster School of Business University of Washington

[email protected]

Jared Stanfield UNSW Business School

UNSW Australia [email protected]

Feng Zhang

David Eccles School of Business University of Utah

[email protected]

This Draft: August 2015

Abstract

This study examines the information that an LBO signals for the target’s industry. We show that LBOs tend to lead merger waves, and in particular predict more LBOs and strategic M&As in the same industry. We establish the valuation implications of this predictability for the LBO target’s industry as a whole. While the industry becomes more valuable following LBOs, surprisingly this is not fully impounded into stock prices at LBO announcement. Instead, as further merger activity occurs, the prices rise, leading to significant abnormal long run returns for peers of an LBO target. We attempt to distinguish between the hypothesis that LBO sponsors simply select into undervalued industries and/or industries that are changing as opposed to causing the changes. We find that our evidence is most consistent with causation, albeit there is some limited evidence of selection. Our study sheds light on the role of LBOs in merger waves as well as the motivation and value-drivers of LBOs. Keywords: leveraged buyouts, mergers and acquisitions, merger waves JEL classification: G14; G34

* We thank Hank Bessembinder, Audra Boone, Karl Lins, Ron Masulis, Konark Saxena, Robert Tumarkin, and Jason Zein, as well as the seminar participants at Australian National University, Deakin University, Fudan University, Massey University, Nanyang Technological University, the University of South Australia, University of Utah, UNSW Australia, and Utah State University for their useful comments. All remaining errors are our own.

Page 2: What does an LBO signal for the target’s industry? · PDF fileWhat does an LBO signal for the target’s industry? Abstract This study examines the information that an LBO signals

What does an LBO signal for the target’s industry?

Abstract

This study examines the information that an LBO signals for the target’s industry. We show that LBOs tend to lead merger waves, and in particular predict more LBOs and strategic M&As in the same industry. We establish the valuation implications of this predictability for the LBO target’s industry as a whole. While the industry becomes more valuable following LBOs, surprisingly this is not fully impounded into stock prices at LBO announcement. Instead, as further merger activity occurs, the prices rise, leading to significant abnormal long run returns for peers of an LBO target. We attempt to distinguish between the hypothesis that LBO sponsors simply select into undervalued industries and/or industries that are changing as opposed to causing the changes. We find that our evidence is most consistent with causation, albeit there is some limited evidence of selection. Our study sheds light on the role of LBOs in merger waves as well as the motivation and value-drivers of LBOs. Keywords: leveraged buyouts, mergers and acquisitions, merger waves JEL classification: G14; G34

Page 3: What does an LBO signal for the target’s industry? · PDF fileWhat does an LBO signal for the target’s industry? Abstract This study examines the information that an LBO signals

1

1. Introduction

A strategic acquisition usually signals synergy gains from combining the operating assets

of the acquirer and target firms. Synergy gains could be triggered by economic shocks such as

changes in technology, regulatory environment, and industry structure (Gort, 1969). The

economic shocks tend to be industry-specific, resulting in “waves” of strategic mergers and

acquisitions within an industry.1 Leveraged buyouts (LBOs) are an important and distinct

subset of mergers and acquisitions (M&As). Despite their importance and distinctiveness, their

potential roles in merger activity remain unexplored. We propose that LBOs signal different

information than strategic M&As; and examine implications of this information for merger

waves and the valuation of the target’s industry. In so doing, we shed light on the motivation and

value-drivers of LBOs more broadly.

Both strategic M&As and LBOs can impact the industry peers of acquisition targets.2

Despite these similarities, strategic M&As and LBOs impact industry peers differently. Unlike

strategic M&As, LBOs involve financial acquirers (mostly private equity funds) instead of

operating acquirers. Hence, LBOs do not signal synergies specific to the acquirer’s operating

assets. Rather, they signal room for improvement within the LBO target firm, which likely

indicates room for improvement in the whole industry. The room for improvement can trigger

acquisition attempts from both financial and strategic acquirers. Therefore, an LBO implies

increased probabilities of both LBOs and strategic M&As in the same industry. On the other

hand, a strategic M&A implies an increased probability of strategic acquisition because

economic shocks tend to be industry-specific (Song and Walkling, 2000). It does not necessarily

1 For example see, Mitchell and Mulherin (1996), Andrade, Mitchell, and Stafford (2001), Rhodes-Kropf and Viswanathan (2004), Harford (2005), Maksimovic, Phillips, and Yang (2012), and Duchin and Schmidt (2013). 2 For example, Servaes and Tomayo (2014) find that hostile strategic takeovers precede significant changes to the target’s industry. Chevalier (1995a, b) study the impact of LBOs on industry competition and pricing policies.

Page 4: What does an LBO signal for the target’s industry? · PDF fileWhat does an LBO signal for the target’s industry? Abstract This study examines the information that an LBO signals

2

imply an increased probability of LBOs because financial acquirers cannot capture synergies

specific to certain operating assets. Because an LBO can signal value potential in an industry for

strategic acquirers, but a strategic acquisition does not necessarily imply the same potential for

an LBO, an LBO is more likely to lead strategic M&As than vice versa. This then implies that

given a merger sequence (wave), LBOs have a greater chance of being the first mover than do

strategic acquirers.

Using a sample of 586 LBOs over the period 1991-2012, we find that an LBO occurring

within an industry significantly increases the likelihood of an industry peer being targeted in

either an LBO or a strategic M&A. On the other hand, the occurrence of a strategic M&A is not

significantly associated with the likelihood of an LBO within the industry. In addition, we find

that aggregate LBO activity significantly predicts aggregate M&A activity, but not vice versa.

Further, we find that LBO activity occurs earlier and LBO sponsors are more likely to be the

first-mover in a given sequence of mergers. On balance, LBOs increase the likelihood of both

LBOs and strategic M&As in the same industry, whereas strategic M&As fail to predict follow-

on LBO activity. The asymmetric predictability is consistent with our conjecture that LBOs

signal different information than strategic M&As.

The increased likelihood of acquisition, which is positive news to shareholders of

potential acquisition targets, implies positive stock returns to industry peers of the LBO target

around the announcement. Consistent with this prediction, we confirm Slovin, Sushka, and

Bendeck (1991)’s finding that industry peers of target firms experience positive abnormal returns

around the LBO announcement.

More importantly, we also find that industry peers of the target firm earn positive

abnormal returns over the twelve months following the LBO announcement. A calendar-time

Page 5: What does an LBO signal for the target’s industry? · PDF fileWhat does an LBO signal for the target’s industry? Abstract This study examines the information that an LBO signals

3

portfolio of LBO industry peers earns a significant alpha of about 1% per month with respect to

the Fama-French (1993) and Carhart (1997) four factors. The abnormal returns are robust over

time and across small and large firms, and are not driven by the “takeover factor” constructed by

Cremers, Nair, and John (2009). This suggests that the spillover effects on industry peers of the

LBO target are much larger than the positive announcement returns to industry peers as

documented in the extant literature. Investors seem to react differently to announcements of

LBOs versus strategic M&As. We find that the same trading strategy involving industry peers of

strategic M&A targets does not produce the same impact among industry peers. The different

investor reactions are also consistent with LBOs signaling different information than strategic

M&As.

There are at least two explanations for the long-run abnormal returns to industry peers of

LBO targets. First, we fail to control for relevant but unknown risk factors. One potential risk is

the uncertainty regarding whether there will be follow-on bids on industry peers of the LBO

target, whether the deal will be withdrawn, what potential changes the private equity firm will

make to the target, and whether they will be value enhancing for the industry as well as the target.

In a market with limits to arbitrage, investors should be compensated for bearing the risk of

uncertain outcomes (see, e.g., Baker and Savasoglu, 2002). Second, the markets fail to fully

incorporate the spillover effects on industry peers into announcement returns. The market

underreaction could be due to investors’ behavioral biases such as overconfidence, optimism,

representativeness or conservatism.3 Consistent with the second explanation, we find that the

long-run abnormal returns are larger for industry peers that are more likely to be acquired after

the LBO.

3 Barberis, Shleifer, and Vishny (1998), Daniel, Hirshleifer, and Subrahmanyam (1998), and Hong and Stein (1999) construct models in which investor psychological biases lead to both under- and over-reaction to financial information.

Page 6: What does an LBO signal for the target’s industry? · PDF fileWhat does an LBO signal for the target’s industry? Abstract This study examines the information that an LBO signals

4

As indicated above, we find that an LBO predicts further merger activity in the target’s

industry, and, more so than for a strategic acquisition, predicts further LBOs. In fact, an LBO is

likely to be early in a sequence of mergers, and deal and target characteristics of acquisitions

following an LBO are different from those coming before it. Further, we find that other firms in

the industry make changes following the LBOs. They are more likely to increase R&D and enter

into alliances following an LBO. They also make governance changes, but rather than improving

governance, we find evidence of defensive changes, including reductions in board independence

and increases in antitakeover provisions and share repurchases.

While the evidence is consistent with private equity (PE) funds inducing changes in the

industry through LBOs, one may argue that changes to industry peers of the LBO target firm are

not caused by the PE funds. Rather, the PE funds could select targets in industries that they

expect to undergo certain changes. For instance, LBOs may indicate the presence of economic

shocks within the same industry and this predictability may be a function of private equity firms’

ability to quickly make acquisitions or their ability to identify undervalued industries as a whole.

In other words, what could appear to be the impact of PE on an industry is actually spuriously

attributed to PE due to its superior speed and selection ability.4

Although it is impossible to completely rule out selection, we conduct tests that suggest

that at least some of the estimated results are caused by the LBO. We do not find evidence that

our results are driven by private equity firms’ ability to quickly make acquisitions following an

economic shock. Specifically, in unreported tests we find that following a regulatory change or

economic shock LBOs are less likely to lead merger sequences and have an insignificant follow-

on impact. Turning to undervaluation, we conduct several other tests. First, if PE firms simply

4 Sørensen (2007) investigates a parallel question in the venture capital market. He finds evidence that private equity firms are able to obtain superior outcomes in the venture capital market by both selecting better deals and adding more value once they are invested in a deal.

Page 7: What does an LBO signal for the target’s industry? · PDF fileWhat does an LBO signal for the target’s industry? Abstract This study examines the information that an LBO signals

5

select into undervalued industries, then follow-on mergers and acquisitions should either be from

outside the industry, or if from within, using cash. However, if these future acquisitions are

caused by real economic change triggered by the private equity firm’s purchase within the

industry, we would expect these future acquisitions to occur both within and between industries

(similar to Mitchell and Mulherin (1996)). We find that LBOs predict follow-on all-stock bids

from within the target’s industry, which is inconsistent with selection.

Next, we examine industry concentration. The causation hypothesis implies that the

abnormal returns of industry peers should be greater in concentrated industries. However, if the

buyout firms are simply selecting into undervalued industries, the value impact coming from the

undervaluation signal would be unrelated to competition within the LBO target industry. We

employ the number of industry peers as well as industry Herfindahl-Hirschman Index (HHI) as

proxies for competition within the industry. We find that announcement returns to industry peers

are negatively (positively) associated with the number of industry peers (HHI). The results are

inconsistent with the selection hypothesis.

The selection hypothesis implies that other institutional investors and hedge funds (often

considered “smart money”) might also be able to determine whether the industry is undervalued

and anticipate future merger activity as well. We find that the number of institutional investors

and the number of hedge funds with holdings in a 4-digit industry significantly increase in the

quarter of and over the four quarters prior to a buyout, and then significantly decrease over the

four quarters after the buyout. In addition, we find that total institutional ownership significantly

decreases over the nine quarters around a buyout. These findings suggest that some, but not all,

“smart” investors enter the industry prior to the LBO and exit later. These investors are able to

capture both the abnormal announcement returns and the long-run abnormal returns to industry

Page 8: What does an LBO signal for the target’s industry? · PDF fileWhat does an LBO signal for the target’s industry? Abstract This study examines the information that an LBO signals

6

peers of the LBO target. This evidence is consistent with the selection hypothesis. Also

potentially in line with the selection hypothesis, we find that firms repurchase more shares from

the market after their industry peers become an LBO target. On balance, our results are

consistent with private equity funds generating changes within industry. Nonetheless, we also

find some, albeit mixed, evidence that LBO acquirers are able to select into undervalued

industries with high future acquisition activity.

Our contributions to the literature are three-fold. First, this paper contributes to the

literature on merger and acquisitions by demonstrating an important link between LBOs and

strategic M&As. Similar to Song and Walkling (2000) but in an LBO context, we find evidence

that LBO activity is a significant determinant of follow-on LBOs and strategic merger activity

within an industry. Our study complements Dittmar, Li, and Nain (2012), who find that

acquisitions of firms that were also targeted by private equity firms tend to perform better. Both

studies suggest that LBOs may have important information for follow-on deals. In addition,

LBOs are not considered in the empirical literature studying the determinants of merger waves or

their effects on firms cited above. We fill the gap and show that LBOs are more likely to be first

movers in merger waves.

This study also enhances our understanding of the broader impacts of private equity

funds beyond those on the individual LBO target firm, and in so doing, provides evidence on the

motivations of the LBOs themselves. Prior studies, which center on individual LBO target firms,

attribute the sources of value creation of LBOs to changes in leverage ratio, investment policy,

operations and corporate governance, as well as value at the eventual sale or IPO.5 Here we

5 See, among many others, Kaplan (1989), Oxman and Yildirim (2008), Guo, Hotchkiss, and Song (2011), Lerner, Sorenson, and Strömberg (2011), Bernstein, Lerner, Sorensen, and Strömberg (2010), Cornelli and Karakas (2012), Acharya, Gottschalg, Hahn, and Kehoe (2013), Axelson, Jenkinson, Strömberg, and Weisbach (2013), Bernstein and

Page 9: What does an LBO signal for the target’s industry? · PDF fileWhat does an LBO signal for the target’s industry? Abstract This study examines the information that an LBO signals

7

broaden the context within which to evaluate LBOs and study what an LBO signals for the

industry of the target. Our finding of positive long-run abnormal returns to industry peers of the

LBO target indicates potential positive impacts from private equity funds. Our additional tests

suggest that these returns are primarily driven by increased acquisition activity in the industry.

In addition, we show that private equity funds not only bring real changes to industry peers but

also are able to move early and often first in a merger wave.

Last but not least, our study contributes to the returns anomaly literature. We find

evidence that, in contrast to non-LBO merger activity, investors do not fully incorporate the

value implications of an LBO within an industry at the LBO announcement. We find a simple

trading strategy, purchasing a portfolio of the industry peers following an LBO, yields positive

and significant long-horizon returns that are not explained by common risk factors, but seem to

vary with ex-post merger activity. Hopefully, our findings can help improve existing asset

pricing models and inspire follow-up studies on the broader impact of private equity funds.

2. Sample Construction and Descriptive Statistics

We identify leveraged buyouts targeting US public companies announced from 1991 to

2012 using Thomson Financial’s SDC database. Our sample period starts in 1991 because we

need to retrieve information on private equity funds/firms from the Preqin database, whose data

coverage begins in 1991. We also require the acquisition to take the form of merger (SDC deal

form M), acquisition of majority interest (AM), or acquisition of assets (AA).6 In addition, we

exclude rumored LBOs and LBOs targeting utility firms (SIC code 4900-4999) and financial

firms (6000-6999). Our final sample consists of 586 LBOs.

Sheen (2013), and Harford and Kolasinski (2013). Excellent reviews of the literature include Metrick and Yasuda (2011) and Eckbo and Thorburn (2013). 6 In unreported tests, our results are robust to dropping deals coded as AA.

Page 10: What does an LBO signal for the target’s industry? · PDF fileWhat does an LBO signal for the target’s industry? Abstract This study examines the information that an LBO signals

8

Table 1 Panel A presents the number of LBOs in each year. There are about ten LBOs

each year in the early 1990s. The number of LBOs per year is much larger in later years, ranging

between 13 LBOs in 2003 and 60 LBOs in 2006.

Table 1 Panel B reports the distribution of the sample LBOs across the Fama-French

(1997) 48 industries. The sample LBOs distribute quite evenly across the industries. There is at

least one LBO in 37 of the 48 industries. Among the 37 industries, the number of LBOs as a

fraction of the firm-years over our sample period ranges between 0.24% for the “Aircraft” and

“Mining” industries and 2.12% for the “Personal Services” industry.

Table 1 Panel C summarizes the characteristics of the LBO deals, the 4-digit SIC industry

peers of the LBO targets, and the associated PE funds. LBO target firms have an average

transaction value of $1.14 billion. The PE firms in our sample offer an average bid premium of

37.5% and the target firms experience average 3-day cumulative abnormal returns (CARs) of

20.6% around the LBO announcement. A substantial number of LBOs precede M&A waves:

about a quarter of the LBO deals are announced during the twenty four months prior to or the

first twelve months of an M&A wave, as defined using the methodology of Harford (2005).

LBO target firms have an average 33.3 industry peers in the 4-digit industry, and a

median of nine industry peers. On average, 5.3% of the industry peers are acquired during the

first year following the LBO, while 12.8% are acquired during the first three years.

On average, there are 1.4 investors for each LBO deal, among which one is a PE investor.

Fifty-three (9.2%) of the sample LBOs have at least one non-PE investor. Management buyouts

(MBOs) account for 28.2% of the sample LBOs. The average lead PE firm is founded 15.5 years

before the LBO, and has $21.1 billion (median of $7.2 billion) under management. On average,

during the preceding three years, the lead PE firm participates in 1.5 LBOs with $6.5 billion

Page 11: What does an LBO signal for the target’s industry? · PDF fileWhat does an LBO signal for the target’s industry? Abstract This study examines the information that an LBO signals

9

involved. The average lead PE fund has $4.0 billion under management, and a sequence order of

7.9. Compared to previous studies such as Metrick and Yasuda (2010), these deals are similar

though slightly larger, given the requirement that the deals be public.

3. LBOs, Merger Likelihood, and Merger Waves

3.1. LBOs and Merger Likelihood

In this subsection, we investigate the influence of LBO activity on future LBO and

strategic M&A activities following the motivation discussed in the introduction. We hypothesize

that the occurrence of an LBO within its industry in the past year increases the likelihood that a

given firm is targeted in an LBO and/or a strategic M&A, and empirically test the hypothesis at

two levels of aggregation. First, Table 2 uses logit analysis at the firm level to test the

likelihood of a firm being targeted in an acquisition in a given year. Second, given the

literature on aggregate and wave activity of M&As (e.g., Harford, 2005) and LBOs (e.g.,

Haddad, Loualiche, and Plosser, 2013), we also examine the aggregate relationship between

LBO activity and M&A activity. Specifically, in Table 3 we test whether aggregate quarterly

LBO activity significantly leads or lags aggregate quarterly M&A activity using vector

autoregression (VAR) analysis with 4 lags.

In Table 2, we find that an LBO announced in the previous year significantly increases

the likelihood that an industry peer becomes the target of an acquisition in model specification

(1). We find this effect is economically significant with an odds ratio of 1.35.7 This effect is

incremental to control variables used in Cremers, Nair, and John (2009) as well as deregulations,

economic shocks, and year fixed effects and with standard errors robust to industry clustering.

7 We calculate this number by exponentiating the estimated log-odds associated with the variable “industry peer targeted in LBO in year t” in model specification (1): exp(0.3013) = 1.35

Page 12: What does an LBO signal for the target’s industry? · PDF fileWhat does an LBO signal for the target’s industry? Abstract This study examines the information that an LBO signals

10

We then decompose the dependent variable into LBO and non-LBO (strategic) targets, and find

that an LBO in the previous year significantly increases the likelihood that an industry peer is the

target of either an LBO or a strategic acquisition (columns (3) and (5)).

Specification (2) shows that the likelihood of an acquisition also increases following a

strategic acquisition within an industry (as discussed in Song and Walkling (2000)). More

notably, we find that this effect is driven by follow-on non-LBO acquisitions. The occurrence of

a non-LBO acquisition is not significantly associated with the likelihood of a follow-on LBO

within the industry (column (4)). Taken as a whole, Table 2 shows that an LBO increases the

likelihood of both an LBO and a non-LBO acquisition within the industry, while a non-LBO

does not increase the likelihood of an LBO.8 We interpret this asymmetric effect as evidence that

the occurrence of an LBO contains information that is different from and uncorrelated with that

provided by a non-LBO.

Performing aggregate VAR analysis in Table 3, we find that aggregate LBO activity in

the previous quarter is associated with significantly higher M&A (and LBO) activity in the

current quarter. This result is consistent with our findings in Table 2. We do not find that the

same holds for non-LBOs in aggregate. Non-LBO merger activity does not significantly predict

future LBO activity, but does predict future M&A activity, again consistent with our firm-level

findings. The results are similar after controlling for macroeconomic variables including real

GDP growth, a GDP price deflator, industrial production, durable consumer goods production,

business equipment production, and the 10-year Treasury yield (the last two columns of Table 3).

Taken together, these results highlight the important impact that LBO activity has (or signals)

8 In unreported robustness tests, inferences remain unchanged if we include an indicator for an LBO acquisition and a non-LBO acquisition within the same specification, if we define industries using SIC 3-digit, Fama French 48 Industries, and Hoberg and Phillips (2010) industry classifications, if we perform our analysis at the industry level, if we follow Gormley and Matsa (2014) and use a linear probability model with interacted year and industry fixed effects, and if we split our analysis by time.

Page 13: What does an LBO signal for the target’s industry? · PDF fileWhat does an LBO signal for the target’s industry? Abstract This study examines the information that an LBO signals

11

about the follow-on merger activity in a given industry. These results are consistent with our

hypothesis, i.e., LBO activity in an industry significantly increases the likelihood of follow-on

acquisition activity, including both LBO and non-LBO acquisitions. This effect differs from the

effects for non-LBO activity found by Song and Walkling (2000). We observe insignificant

effects for aggregate LBO activity in the two, three, or four quarters prior. One potential reason

is that aggregating over industries could lower the time-series variation of the number of deals,

and consequently attenuates or masks the effects of LBO/M&A activity from the farther past on

future activities at the industry level. In what follows, we further investigate the position of

LBOs within merger sequences, and their effects on industry peers.

3.2. Are LBOs First Movers in Merger Sequences?

We further investigate potential impacts of an LBO on industry peers of the LBO target

by examining whether LBOs lead general merger and acquisition activity in merger sequences.

In Table 4, we define a merger sequence as beginning (ending) with an announced acquisition in

a 4-digit SIC industry if there has not been an acquisition announced in the same industry in the

previous (following) 365 days between 1990 and 2012. As seen in Table 4, Panel A, we classify

666 merger sequences (containing at least 2 deals). On average, each sequence consists of 7.3

deals, and lasts 2.5 years. There are 984 single deals not contained in any merger sequence,

representing 18% of the 5,436 deals in our sample. The average location of an LBO within a

given merger sequence is 31% through the sequence, indicating that LBOs tend to occur earlier

in a sequence.

Table 4, Panels B and C report evidence consistent with LBOs occurring earlier in

merger sequences and being more likely to be first-movers within a given merger sequence.

Page 14: What does an LBO signal for the target’s industry? · PDF fileWhat does an LBO signal for the target’s industry? Abstract This study examines the information that an LBO signals

12

Panel B presents regression results using public targets while Panel C presents similar analysis

(without the accompanying control variables) using both public and private targets. Specification

(1) of both panels, where the dependent variable is the logarithm of the sequence number, shows

that LBOs occur significantly earlier in a given merger sequence, whether we analyze public

targets or all targets. Specification (1) includes the 984 single deals not within a merger

sequence. To ensure our results are not being driven by LBOs being more likely to occur alone,

we exclude the 984 deals in specification (2). The results remain consistent across both panels.

In addition, in untabulated robustness analysis we find that LBOs are no more likely to occur

alone than other acquisitions.

Specifications (3)-(4) reports logit analysis on similar samples to specifications (1)-(2),

where the dependent variable is the first mover indicator, which takes the value of one for the

first deal of a merger sequence. We present evidence that LBOs are significantly more likely to

occur first within a merger sequence (despite the fact that the number of LBOs is roughly 10% of

the total sample). Economically, an LBO is roughly 59% (odds ratio of 1.59) more likely to be a

first mover in a sequence of mergers. The evidence remains unchanged whether we analyze

public targets only or include private targets as well. If sequences of mergers containing LBOs

are significantly shorter than those not containing LBOs, the above findings may be mechanical.

However, in untabulated results we find the opposite to be the case: sequences containing LBOs

contain significantly more acquisitions. Taken together, the evidence of Table 4 is consistent

with LBOs being more likely to be first-movers preceding other acquisition activity.

Page 15: What does an LBO signal for the target’s industry? · PDF fileWhat does an LBO signal for the target’s industry? Abstract This study examines the information that an LBO signals

13

3.3. Characteristics of Deals before versus after LBOs

The results discussed so far indicate that LBOs affect or signal follow-on merger activity

in a given industry. That is, LBOs convey information about potential synergies of acquiring

industry peers of the LBO target. More aware of the magnitude of potential synergy gains,

managers of industry peers of the LBO target are expected to bargain for higher bid premiums

when they become an acquisition target, which in turn will result in higher target firm

announcement returns, lower acquirer firm announcement returns, and higher likelihoods of deal

completion.

In Table 5 we examine differences of the characteristics between acquisitions following

an LBO versus those preceding the LBO or deals not surrounding LBOs. Panel A reports the

difference in each of the deal characteristics between the non-LBO acquisitions following the

first LBO of a given merger sequence and those preceding the first LBO. Any differences in

Panel A could be the effect of the LBO, or simply be the result of comparing deals early in a

sequence to late in a sequence. In order to address the concern of deal location in a sequence

containing an LBO and to compare these differences to deals in sequences not containing an

LBO, we randomly choose a deal as a pseudo-LBO in each sequence using the mean and

standard deviation of the location of the first LBO in sequences containing an LBO. We then

compute the differences of the deal characteristics following and preceding the pseudo-LBOs,

tabulate the number of deals in each group, and calculate the significance of the differences. We

repeat this procedure 500 times and report the averages in Table 5, Panel B. Panel C reports the

differences between Panels A and B as well as the difference-in-differences.

Table 5, Panel A, shows that deals following LBOs have higher bid premiums, higher

target firm announcement returns, lower acquirer announcement returns, and higher likelihoods

Page 16: What does an LBO signal for the target’s industry? · PDF fileWhat does an LBO signal for the target’s industry? Abstract This study examines the information that an LBO signals

14

of deal consummation. The differences in the four characteristics are all statistically significant.

These differences are not due to being late in a given merger sequence: They all become

insignificant when we replace the real LBOs with pseudo LBOs in Panel B. The difference-in-

differences, which compare the difference between deals before and after real LBOs and the

deals before and after pseudo-LBOs, remain positive and statistically significant for target

announcement returns and deal completion rate, although they are no longer statistically

significant for bid premium and acquirer announcement returns (Panel C). On balance, the

results indicate that the occurrence of an LBO not only significantly increases the likelihood of

follow-on acquisitions, but that the follow-on deals are significantly different from those

following non-LBOs. This is consistent with LBOs signaling different information to the market

than strategic M&As.

Taking Tables 2-5 together, these results suggest that LBO activity has follow-on effects

for industry peers and that these effects are incremental to those previously found following non-

LBO acquisitions. These in turn have important value implications for industry peers of the

LBO target, which we investigate in the next section.

4. Value Implications of LBOs for Industry Peers of the LBO Target

4.1. Short- and Long-Run Returns to Industry Peers of the LBO Target

In the previous section, we find that firms are significantly more likely to be targeted in

an acquisition if an industry peer has been targeted in an LBO in the previous year. This

increased likelihood of acquisition implies positive information to the industry peers of LBO

targets.

Page 17: What does an LBO signal for the target’s industry? · PDF fileWhat does an LBO signal for the target’s industry? Abstract This study examines the information that an LBO signals

15

We first present the short-run abnormal returns to industry peers of LBO target firms in

Table 6. Following Brown and Warner (1980), we compute abnormal returns based on the

market model. We observe that the portfolio of industry peers has positive and significant

abnormal returns in the 3-day window surrounding the announcement (and also in the month

following) the LBO. Economically, the average portfolio firm has an abnormal return of 0.4%

around the announcement and 0.7% in the month following the LBO announcement. The results

confirm Slovin, Sushka, and Bendeck’s (1991) finding of positive abnormal returns to industry

peers of the LBO target over the few days around the LBO announcement, but using our more

recent and expanded sample. In addition, our findings indicate stock return drift for industry

peers of the LBO target following the LBO announcement.

In an efficient market, all relevant information should be incorporated into stock prices

around the short window surrounding the LBO announcement. However, extensive evidence

shows that investors might under- or over-react to information (see Footnote 3). We investigate

the long-run impact of LBOs on industry peer returns in Table 7. At the beginning of each

month from January 1991 to December 2012, we form a portfolio of firms where at least one

target in our sample of LBOs is in those firms’ 4-digit SIC industry and was announced during

the preceding 12 months.9 The industry peers are excluded from the portfolio once the LBO deal

is withdrawn.10 Thus, we are tracking the monthly returns of firms in an industry that has

experienced an LBO within the last 12 months. We find that the portfolio of industry peers earns

9 We employ the calendar time portfolio approach to investigate long-run abnormal returns to industry peers because Bessembinder and Zhang (2013) find that the traditional buy-and-hold abnormal return approach cannot effectively control for important differences in firm characteristics between the event firm and its matching firm. In addition, while we feel that the effects of LBOs on industry peers will be more pronounced using narrower industry definitions, our inferences remain unchanged when we define industries at the SIC 3-digit or Fama French 48 level. 10 Thirty-five percent (35%) of our sample LBOs are not successfully completed. Including industry peers in the monthly portfolio after deal withdrawal does not alter our results.

Page 18: What does an LBO signal for the target’s industry? · PDF fileWhat does an LBO signal for the target’s industry? Abstract This study examines the information that an LBO signals

16

equal-weighted (value-weighted) returns of 1.76% (1.35%) per month over the 12 months

following the LBO.

As can be seen in Table 7, Panel B, specifications (1) – (2), the industry peer portfolio

generates an alpha of 0.99% per month on an equal-weighted basis after controlling for the four

risk factors (MKT, SMB, HML, and UMD) constructed by Fama and French (1993) and Carhart

(1997). The value-weighted alpha is 0.58%. The estimated alphas are statistically significant at

the one percent level with associated t-statistics of 4.24 and 3.55, respectively. The t-statistics

are much greater than the cut-off value of three suggested by Harvey, Liu, and Zhu (2014). This

significantly alleviates the concern that our findings are the result of data mining.

The positive long-run abnormal returns are robust to a battery of tests. Cremers, Nair,

and John (2009) find evidence that a “takeover factor”, constructed by buying (selling) firms

with high (low) takeover likelihood, generates an abnormal return that represents higher risk. In

specifications (3) – (4) we explicitly control for the takeover factor and continue to find

significant abnormal returns. Finally in Panel C, this effect is economically and statistically

significant across both small and large firms. It is stronger in earlier years (1991-2001) but

remains both economically and statistically significant in later years (2002-2012) in the equal

weighted portfolio. In unreported analysis, we do not find evidence of a reversal of this effect

over the second or the third year following the LBO announcement. In fact, we find significantly

positive, albeit smaller, alphas for the portfolio of industry peers during those years.

Given that the occurrences of LBOs and non-LBOs increase the likelihood of future

acquisitions within an industry, it is possible the long-run returns we find for industry peers of

LBO targets in Table 7 are also applicable to industry peers of non-LBO targets. Table 8

presents analysis similar to Table 7, but uses the occurrence of non-LBO mergers, rather than

Page 19: What does an LBO signal for the target’s industry? · PDF fileWhat does an LBO signal for the target’s industry? Abstract This study examines the information that an LBO signals

17

LBOs, to form industry portfolios. We find significantly weaker estimated alphas: with a

significant abnormal return of 31 basis points for the equally-weighted portfolio, and an

insignificant 9 basis points in the value-weighted portfolio.

Taking the results of Table 7 and Table 8 together, we find evidence consistent with LBO

activity influencing long-run industry valuation. Specifically, we find that the market does not

fully impound the long-run value changes of LBOs to industry peers around the announcement

of the LBO. The abnormal returns are robust and can be captured by a simple trading strategy.

However, we do not find the same effect for non-LBO M&A activity. The results suggest that

investors react differently to announcements of LBOs and strategic M&As. This is consistent

with the conclusion that they signal different information to the market.

4.2. Increased Likelihood of Acquisition and Stock Returns to Industry Peers of LBO Targets

The results up to this point show (1) that the occurrence of an LBO in an industry

significantly increases the likelihood of M&A activity within that industry; and (2) that LBO

activity is associated with short- and long-run abnormal returns to industry peers. As noted

before, the observed increased likelihood of being acquired is value-enhancing and thus could

potentially be responsible for the abnormal returns to industry peers of the LBO target. In this

subsection, we investigate the association between the abnormal returns and the likelihood of

acquisition ex post. Specifically, we divide the sample LBOs into two groups based on the

fraction of industry peers of the LBO target that become acquisition targets over the following

one or three years.

Table 9 reports short-run abnormal returns (Panel A) and long-run abnormal returns

(Panel B) split into sub-samples by ex-post measures of increased merger activity. In Table 9,

Page 20: What does an LBO signal for the target’s industry? · PDF fileWhat does an LBO signal for the target’s industry? Abstract This study examines the information that an LBO signals

18

Panel A, we find that short-run industry peer CARs are actually lower when the LBO precedes

high merger activity, whether as defined using the method in Harford (2005), or by realized

probability of acquisition over 1 or 3 years.

In Table 9, Panel B, we repeat the comparisons in Panel A but use the estimated long-run

Jensen’s alpha. As opposed to the short-run results of Panel A, we find evidence that the

estimated positive long-run abnormal returns were centered in industries with high ex-post

merger activity, or in LBOs that led merger waves. We also examine low and high loan spread

periods and find that periods of lower loan spreads obtained higher abnormal returns consistent

with heightened takeover activity by reputable firms and debt-usage as found by Harford (2005),

Demiroglu and James (2010) and Axelson, Jenkins, Strömberg, and Weisbach (2013). Table 9

therefore provides some evidence regarding the source of our observed abnormal long-run

returns, specifically evidence that they are centered in industries with the largest amount of

subsequent merger activity.11 Thus, prices appear to underreact at the LBO announcement and

then increase as merger activity follows, producing the observed long-run abnormal returns.

5. Selection or Causation

Private equity funds may influence an industry through changes in financial policy—

especially leverage ratios, governance, pricing, or the structure and strategic relationships within

the industry itself. Alternatively, private equity firms may simply anticipate or be able to react

quickly to future changes within the industry or identify undervalued industries and select into

these industries. This ability to spur changes within an industry or select into an industry may

differ based on the composition of the investment syndicate or the characteristics of the private

11 In untabulated analysis, we find evidence that our estimated long-run abnormal returns significantly increase when an industry peer is targeted in a management buyout (MBO) and when the LBO deal does not involve non-PE firms. We do not find any evidence that fund size, age, or experience influences the estimated long-run returns.

Page 21: What does an LBO signal for the target’s industry? · PDF fileWhat does an LBO signal for the target’s industry? Abstract This study examines the information that an LBO signals

19

equity firms. For example, more experienced firms arguably can influence more changes within

a firm and therefore its industry or identify undervalued industries more accurately. A

management-led buyout potentially signals the management’s belief of potential improvements

within an industry or that the industry is undervalued.

As discussed earlier, Sørensen (2007) investigates the source of gains by venture capital

firms, disentangling potential benefits through selection (obtaining better deals) or through

value-improvement (making more improvements after the deal is made). We similarly wish to

investigate whether the increased likelihood of acquisition and the abnormal returns following an

LBO were caused by the LBO (as argued by Slovin, Sushka, and Bendeck, 1991 and Bernstein,

Lerner, Sorensen, and Strömberg, 2010) or whether LBO firms are simply very good at selecting

into industries that are undervalued or will experience these changes regardless of their

involvement. To disentangle the different (though not necessarily exclusive) effects of our

selection and causation, we perform a series of tests in the rest of this section.

5.1. Economic Changes to the LBO Target Industry

The selection hypothesis primarily centers on undervaluation, while the causation

hypothesis states that LBOs will bring real changes to industry peers of the LBO target. As a test

of the causation hypothesis, we examine post-LBO leverage ratios, strategic actions and changes

in corporate governance of industry peers of the LBO target. We focus on these three aspects

because prior studies show that private equity funds can impact target firms through financial

engineering (taking advantage of tax shields with high leverage), making operational changes,

and changing the governance structure of the firm (see Footnote 5 for related studies).

Page 22: What does an LBO signal for the target’s industry? · PDF fileWhat does an LBO signal for the target’s industry? Abstract This study examines the information that an LBO signals

20

In Table 10, Panel A, we do not find evidence that peer firms significantly change cash

holdings or leverage ratios in the three years following the LBO announcement.12 Thus, they do

not mimic the changes made by LBO targets or converge to their characteristics. In Table 10,

Panel B, we find evidence that the industry peers of targets of LBOs increase R&D, and are

significantly more likely to undertake an alliance or a strategic alliance in the three years

following the LBO announcement. This provides some support to the conclusion that LBOs

shock the competitive nature of an industry (as found by Chevalier, 1995a;b), which causes

rivals to undergo strategic changes, either through higher R&D investment or through alliances

with other firms in addition to a heightened probability of being targeted in an acquisition.

However, we do not find any evidence that the average peer of an LBO target is more likely to

have a spinoff following an LBO. Our findings are consistent with Bernstein, Lerner, Sorensen,

and Strömberg (2010), who, in a cross-country study, find evidence that LBOs cause rivals to

mimic changes made by the LBO target.

LBOs can spur industry peers to mimic LBO targets and improve governance. In Table

10, Panel C, we find evidence that the opposite may be the case. Analyzing several measures of

the change in corporate governance—board size, the proportion of independent directors, the

governance index (G Index; Gompers, Ishii, and Metrick, 2003), the entrenchment index (E

Index; Bebchuk, Cohen, and Farrell, 2009), and total CEO pay—we find that the governance of

peer firms of LBO targets generally worsens following an LBO. Specifically, board size, the G

Index, and the E Index significantly increase while boards become significantly less independent

12 All results of Table 10 are robust, though in some cases with reduced significance, when we define the LBO indicator as one if an industry peer has been targeted in the previous year.

Page 23: What does an LBO signal for the target’s industry? · PDF fileWhat does an LBO signal for the target’s industry? Abstract This study examines the information that an LBO signals

21

following an industry peer being targeted in an LBO.13 This provides some evidence that rather

than mimicking the improved governance of an industry peer targeted in an LBO, governance

worsens at peer firms, perhaps in response to the increased likelihood of acquisition.

Share repurchases usually raise stock prices and thus help deter potential takeover

attempts (Bagwell, 1991; Sinha, 1991; and Dittmar, 2000). Therefore, we investigate whether

peer firms increase share repurchases in response to the increased likelihood of acquisition after

an LBO announcement in the same industry. Table 10, Panel D demonstrates evidence

consistent with this prediction: firms significantly increase share repurchases following the LBO

of an industry peer. The results in Panels C-D of Table 10 suggest that industry peers of the

LBO target are aware of and actually respond to the increased likelihood of acquisition.

Table 11 presents difference-in-difference analysis similar to Table 5, and we find that

deals following LBOs are significantly more likely to involve industries with strong customer-

supplier connections. Again, this is supportive of the LBO causing strategic changes within an

industry, which causes follow-on acquisitions along the industry supply chain. Additionally,

while the results of several other control variables suggest deals following LBOs are

significantly different than those that do not, they do not appear to be driven by Servaes and

Tamayo (2014)’s findings of the impacts of hostile takeover attempts on industry peers.

Specifically, we find that targets of deals following LBOs have lower leverage, worse accounting

performance, and are less likely to be hostile. This result also supports the findings of Table 10

and suggests that industry peers (and follow-on targets) do not mimic the increased leverage of

the LBO.

13 These findings are in contrast to those found by Oxman and Yildirim (2009), which find governance improves in industries following LBOs. These differences are likely the result of a different sample period (1992-2006) and industry definition (2-digit SIC industries).

Page 24: What does an LBO signal for the target’s industry? · PDF fileWhat does an LBO signal for the target’s industry? Abstract This study examines the information that an LBO signals

22

In summary, Table 10 and Table 11 provide evidence that LBO activity spurs changes in

industry peers including increased R&D, an increased likelihood of undertaking alliances with

other firms, worsening governance, and increased likelihood of acquisitions involving suppliers

and customers. However, firms do not appear to change leverage in response to a peer being

targeted in an LBO. In Tables 5 and 11, we report evidence that deals following an LBO have

significantly different characteristics than those preceding it. Altogether, these results support

the causation hypothesis rather than selection.

5.2. Diversifying Deals and Method of Payment

We also test the selection and causation hypotheses by analyzing whether the likelihood

of a given deal being within-industry and all stock-financed is related to an industry peer of the

target being targeted in an LBO in the previous year. We similarly test whether the likelihood of

a bidder of an announced all stock-financed deal is related to an industry peer of the bidder being

targeted in an LBO in the previous year. If a private equity firm were selecting into an

undervalued industry, we would expect acquirers from within the industry to be less likely to

acquire the industry peers of an LBO target in all stock-financed deals. Similarly, we would

expect industry peers of a target LBO to be less likely to finance a deal with all stock.

Table 11 presents univariate evidence and Table 12 presents multivariate evidence

contrary to the selection hypothesis: given a firm is the target of an acquisition, the likelihood

that the acquisition is within-industry and all stock-financed is significantly higher if an LBO

occurred within that industry. In addition, as found in Table 12, specification (2), industry peers

of LBO targets are more likely to acquire another firm in an all stock-financed deal. Given this

analysis only requires public acquirers rather than public targets, the sample is significantly

Page 25: What does an LBO signal for the target’s industry? · PDF fileWhat does an LBO signal for the target’s industry? Abstract This study examines the information that an LBO signals

23

larger than the analysis of public targets found in specification (1). Economically, the

occurrence of an LBO within a given industry increases the likelihood of a given acquisition

being within-industry and all stock-financed is associated with an odds ratio of 1.54 and

increased the likelihood that an industry peer acquires another firm in an all stock-financed deal

by an odds ratio of 1.20. These results are at odds with the selection hypothesis.

5.3. Industry Competition and Abnormal Returns to Industry Peers of the LBO Target

In addition, we test whether industry competition relates to the estimated abnormal

returns to industry peers of the LBO target in Table 13, Panels A-C. The selection hypothesis

predicts that the abnormal returns, both short- and long-run, will only be related to valuation of

industry peers of the LBO target, but will not be related to industry-level competition.

We find that industry competition (as measured by the natural log of the number of

industry peers) is negatively associated with short-run abnormal returns and industry

concentration (as measured by HHI) is positively associated with short-run abnormal returns in

Panels A and B. The results continue to hold after controlling for valuation (market-to-book ratio)

of the LBO target industry. Industry peers in concentrated industries are more likely to face the

strategic effects of an LBO. The market impounds this information on announcement. In Table

13, Panel C, we find that our long-run abnormal returns are centered in less-concentrated

industries (higher number of industry peers and a lower HHI index). Taken together, these

results are inconsistent with the selection hypothesis.14

14 In untabulated results, we find the characteristics of private equity funds (reported in Table 1) are not significantly associated with announcement CARs to industry peers of the LBO target, with only one exception. Private equity fund size is significantly related to positive abnormal short-run CARs for industry peers. This is consistent with the market believing that a large fund can cause (or identify) a greater impact within an industry. It does not help us disentangle the causation hypothesis versus the selection hypothesis.

Page 26: What does an LBO signal for the target’s industry? · PDF fileWhat does an LBO signal for the target’s industry? Abstract This study examines the information that an LBO signals

24

5.4. Changes in Institutional Investor Ownership around LBOs

Furthermore, we investigate the holdings and trading behavior of other institutional

investors around the LBO. The purchase of an entire publicly traded firm is significantly more

costly and risky than purchasing the shares of that firm. As such, the gains of this acquisition

should be great enough to outweigh the associated costs. If private equity firms are selecting

into undervalued industries that will undergo value-enhancing changes in the future, other

“smart-money” investors may be able to do the same.

We test this prediction of the selection hypothesis explicitly in Table 14 using changes in

both the holdings and number of institutional investors (in specifications (1) and (3)) and hedge

funds (specifications (2) and (4)) that have invested in the target industry in the 4 quarters before

and the 4 quarters after the LBO is announced and control for industry and time fixed effects.

We find that the number of institutional investors in the industry significantly increases in the

quarter of and the 4 quarters prior to the LBO announcement. The number of hedge funds also

significantly increases in the quarter of the LBO announcement. Both the number of institutional

investors and the number of hedge fund investors significantly decrease over the 4 quarters after

the LBO. In addition, we find that aggregate institutional ownership significantly decreases

around the LBO announcement. It does appear that some but not all institutional investors (or

hedge funds) enter the industry before the LBO announcement and exit later. These investors are

able to capture the estimated average abnormal announcement returns and abnormal long-run

returns on average that we find. Taken together the results of Table 14 provide some, albeit

mixed, evidence in support of the selection hypothesis.

Page 27: What does an LBO signal for the target’s industry? · PDF fileWhat does an LBO signal for the target’s industry? Abstract This study examines the information that an LBO signals

25

5.5. Share Repurchases by Industry Peers of the LBO Target

As discussed previously, we investigate share repurchases by industry peers of the LBO

target. The selection hypothesis predicts that they will repurchase under-valued shares. In Table

10, Panel D, we observe that firms tend to repurchase more shares from the market after their

industry peers become an LBO target. The results are potentially consistent with the selection

hypothesis. However, given our prior discussion of these results, increased share repurchases

may also be due to firms recognizing and attempting to shield themselves from increased

takeover activity.

5.6. Deals following Deregulation or Economic Shock

Finally, an alternate explanation of our results is that they are being driven by private

equity firms’ ability to quickly react to regulatory changes or economic shocks. To determine

whether this is the case, we analyze the follow-on impact of LBOs on other acquisitions as well

as the first-mover analysis following a regulatory change or economic shock. In untabulated

tests, we do not find that LBOs that follow a regulatory change or economic shock significantly

predict other acquisitions and nor lead merger sequences. This reduces the likelihood that

private equity firms reacting quickly to changes rather than causing changes drive our results. In

addition, if our results are due to the quick response of private equity firms, it is unclear how the

LBO, rather than the economic shock, would influence the abnormal returns of industry peers.

On balance, the test results provide supportive evidence for the causation hypothesis, and

much weaker evidence consistent with the selection hypothesis. We conclude that PE funds

bring economic changes to the target industry and do not simply select into industries where they

Page 28: What does an LBO signal for the target’s industry? · PDF fileWhat does an LBO signal for the target’s industry? Abstract This study examines the information that an LBO signals

26

expect changes to occur or quickly respond to economic shocks. However, some PE activity is

driven by selection.

6. Conclusion

Funds under management by private equity firms have increased substantially over the

last 20 years, and with this increase, comes more leveraged buyouts. LBOs have been shown to

substantially impact the operations of the target firm. In doing so, they also have implications for

the target’s industry peers. We document that LBOs are often a precursor to an acquisition wave

within an industry and that if anything, the stock market underreacts to this fact. Long-run

abnormal returns to LBO target industry peers are positive and significant and appear to be

primarily driven by increased merger activity in the industry.

In addition to documenting the subsequent merger activity and the returns anomaly it

generates, we also document changes made by industry peers following an LBO. Rather than

changing financial characteristics (leverage, cash holdings), they make real investment changes

such as increasing R&D and entering into alliances. If anything, their governance deteriorates,

which can be best explained as adopting a more defensive posture, likely in response to the

increased likelihood of acquisition activity.

We explore whether private equity firms are simply good at selecting into undervalued

industries (or industries poised for changes) or actively cause the changes in the industry. The

two are not mutually exclusive, and while selection partly motivates LBOs, we find more support

for the causation hypothesis as being representative of LBO activity. This has the wider

implication that LBOs in an industry are motivated by the potential for real value improvements

in the targets, not simply as opportunistic selection of undervalued targets.

Page 29: What does an LBO signal for the target’s industry? · PDF fileWhat does an LBO signal for the target’s industry? Abstract This study examines the information that an LBO signals

27

References:

Acharya, Viral V., Oliver F. Gottschalg, Moritz Hahn, and Conor Keho, 2013. Corporate governance and value creation: Evidence from private equity. Review of Financial Studies 26, 368-402.

Andrade, Gregor, Mark Mitchell, and Erik Stafford, 2001. New evidence and perspectives on mergers. Journal of Economic Perspectives 15, 103-120.

Axelson, Ulf, Tim Jenkinson, Per Strömberg, and Michael S. Weisbach, 2013. Borrow cheap, buy high? The determinants of leverage and pricing in buyouts. Journal of Finance 68, 2223-2267.

Bagwell, Laurie S., 1991. Share repurchase and takeover deterrence. RAND Journal of Economics 22, 72-88.

Banyi, Monica L., Edward A. Dyl, and Kathleen M. Kahle, 2008. Errors in estimating share repurchases. Journal of Corporate Finance 14, 460–474.

Barberis, Nicholas, Andrei Shleifer, and Robert Vishny, 1998. A model of investor sentiment. Journal of Financial Economics 49, 307–345. Baker, Malcolm, and Serkan Savasoglu, 2002. Limited arbitrage in mergers and acquisitions. Journal of Financial Economics 64, 91–115. Bebchuk, Lucian A., Alma Cohen, and Allen Ferrell, 2009. What matters in corporate governance?. Review of Financial Studies 22, 783–827. Bernstein, Shai, Josh Lerner, Morten Sørensen, and Per Strömberg, 2010. Private equity and industry performance. NBER Working Paper No. 15632.

Bernstein, Shai, and Albert Sheen, 2013. The operational consequences of private equity buyouts: Evidence from the restaurant industry. Rock Center for Corporate Governance at Stanford University Working Paper No. 156.

Bessembinder, Hank, and Feng Zhang, 2013. Firm characteristics and long-run stock returns after corporate events. Journal of Financial Economics 109, 83–102.

Betton, Sandra, B. Espen Eckbo, and Karin S. Thorburn, 2008. Corporate Takeovers. In: Handbook of Corporate Finance: Empirical Corporate Finance, Vol. 2, Chapter 15, pp. 291-430.

Brown, Stephen J., and Jerold B. Warner, 1980. Measuring security price performance. Journal of Financial Economics 8, 205–258. Cai, Jie, Moon H. Song, and Ralph A. Walkling, 2011. Anticipation, acquisitions, and bidder returns: Industry shocks and the transfer of information across rivals. Review of Financial Studies 24, 2242-2285.

Carhart, Mark M., 1997. On persistence in mutual fund performance. Journal of Finance 52, 57–82.

Chevalier, Judith A., 1995a. Capital structure and product-market competition: Empirical evidence from the supermarket industry, American Economic Review 85, 415-435.

Page 30: What does an LBO signal for the target’s industry? · PDF fileWhat does an LBO signal for the target’s industry? Abstract This study examines the information that an LBO signals

28

Chevalier, Judith A., 1995b. Do LBO supermarkets charge more? An empirical analysis of the effects of LBOs on supermarket pricing. Journal of Finance 50, 1095-1112.

Cornelli, Francesca and Oguzhan Karakas, 2012. Corporate governance of LBOs: The role of boards. Unpublished working paper, London Business School.

Cremers, K. J. Martijn, Vinay B. Nair, and Kose John, 2009. Takeovers and the cross-section of returns. Review of Financial Studies 22, 1409–1445.

Daniel, Kent, David Hirshleifer, and A. Subrahmanyam, 1998. Investor psychology and security market under- and overreactions. Journal of Finance 53, 1839–1885.

Demiroglu, Cem, and Christopher M. James, 2010. The role of private equity group reputation in LBO financing. Journal of Financial Economics 96, 306-330.

Dittmar, Amy, 2000. Why do firms repurchase stock? Journal of Business 73, 331-355.

Dittmar, Amy, Di Li, and and Amrita Nain, 2012. It pays to follow the leader: Acquiring targets picked by private equity. Journal of Financial and Quantitative Analysis 47, 901–931.

Duchin, Ran and Breno Schmidt, 2013. Riding the merger wave: Uncertainty, reduced monitoring, and bad acquisitions. Journal of Financial Economics 107, 69-88.

Eckbo, B. Espen, 1983. Horizontal mergers, collusion and stockholder wealth. Journal of Financial Economics 11, 241–273.

Eckbo, B. Espen, and Karin S. Thorburn, 2013. Corporate Restructuring. Foundations and Trends in Finance, Forthcoming.

Fama, Eugene, and Kenneth French, 1993. Common risk factors in the returns on stocks and bonds. Journal of Financial Economics 33, 3–56.

Fama, Eugene, and Kenneth French, 1997. Industry costs of equity. Journal of Financial Economics 43, 153–193.

Gompers, Paul, Joy Ishii, and Andrew Metrick, 2003. Corporate governance and equity prices. Quarterly Journal of Economics 118, 107–155.

Gormley, Todd A., and David A. Matsa, 2014. Common errors: How to (and not to) control for unobserved heterogeneity. Review of Financial Studies 27, 617-661.

Gort, Michael, 1969. An economic disturbance theory of mergers. Quarterly Journal of Economics 83, 624-642.

Guo, Shourun, Edith S. Hotchkiss, and Weihong Song, 2011. Do buyouts (still) create value? Journal of Finance 66, 479-517.

Haddad, Valentin, Erik Loualiche, and Matthew C. Plosser, 2013. Buyout activity: The impact of aggregate discount rates. FRB of New York Staff Report No. 606.

Harford, Jarrad, 2005. What drives merger waves?. Journal of Financial Economics 77, 529–560.

Page 31: What does an LBO signal for the target’s industry? · PDF fileWhat does an LBO signal for the target’s industry? Abstract This study examines the information that an LBO signals

29

Harford, Jarrad and Adam Kolasinski, 2013. Do private equity returns result from wealth transfers and short-termism? Evidence from a comprehensive sample of large buyouts. Management Science 60, 888-902.

Harvey, Campbell, Yan Liu, and Heqing Zhu, 2014. … and the cross-section of expected returns. Unpublished working paper, Duke University.

Hoberg, Gerard, and Gordon Phillips, 2010. Product market synergies and competition in mergers and acquisitions. Review of Financial Studies 23, 3773–3811.

Hong, Harrison, and Jeremy Stein, 1999. A unified theory of underreaction, momentum trading, and overreaction in asset markets. Journal of Finance 54, 2143-2184.

Kaplan, Steven N., 1989. The effects of management on operating performance and value. Journal of Financial Economics 24, 217-254.

Lerner, Josh, Morten Sorensen, and Per Strömberg, 2011. Private equity and long-run investment: The case of innovation. Journal of Finance 66, 445-477.

Maksimovic, Vojislav, Gordon Phillips, and Liu Yang, 2013. Private and public merger waves. Journal of Finance 68, 2177-2217.

Metrick, Andrew, and Ayako Yasuda, 2011. Venture capital and other private equity: A survey. European Financial Management 17, 619–654.

Mitchell, Mark L., and J. Harold Mulherin, 1996. The impact of industry shocks on takeover and restructuring activity. Journal of Financial Economics 41, 193-229.

Oxman, Jeffrey and Yildiray Yildirim, 2008. Governance effects of LBO events. Unpublished working paper, Syracuse University.

Servaes, Henri and Ane Tamayo, 2014. How do industry peers respond to control threats? Management Science 60, 380-399.

Sinha, Sidharth, 1991. Share repurchase as a takeover defense. Journal of Financial and Quantitative Analysis 26, 233-244.

Slovin, Myron B., Marie E. Sushka, and Yvette M. Bendeck, 1991. The intra-industry effects of going-private transactions. Journal of Finance 46, 1537-1550.

Song, Moon H., and Ralph A. Walkling, 2000. Abnormal returns to rivals of acquisition targets: A test of the “acquisition probability hypothesis”. Journal of Financial Economics 55, 143–171.

Sørensen, Morten, 2007. How smart is smart money? A two-sided matching model of venture capital. Journal of Finance 62, 2725-2762.

Page 32: What does an LBO signal for the target’s industry? · PDF fileWhat does an LBO signal for the target’s industry? Abstract This study examines the information that an LBO signals

30

Table 1: Summary statistics Panel A presents the number of LBOs by year, while Panel B presents the number of LBOs by industry. Panel C summarizes characteristics of the sample LBOs and characteristics of the acquiring private equity funds/firms. Transaction value is the size of the acquisition reported in billions of dollars. Bid premium is the ratio of the final offer price to the target stock price four weeks prior to the announcement date minus one. Target 3-day announcement CARs are cumulative abnormal returns over days (-1, +1) around the LBO announcement based on the market model (Brown and Warner, 1980), with beta estimated over the period from 42 to 252 days prior to the announcement date. An LBO leads a merger wave if it is announced over the 24 months prior to or over the first 12 months after the beginning of a merger wave in the same industry as defined by the methods of Harford (2005). Non-PE deal is a dummy variable that takes the value of one if non-private equity acquirers participate in the LBO, and zero otherwise. Our sample includes LBOs targeting public US firms announced between 1991 and 2012. Panel A: Number of LBOs by year Year N 1991 8 1992 11 1993 8 1994 9 1995 14 1996 17 1997 30 1998 35 1999 47 2000 52 2001 18 2002 22 2003 13 2004 22 2005 32 2006 60 2007 54 2008 23 2009 17 2010 38 2011 24 2012 32 Total 586

Page 33: What does an LBO signal for the target’s industry? · PDF fileWhat does an LBO signal for the target’s industry? Abstract This study examines the information that an LBO signals

31

Panel B: Number of LBOs by target industry

Fama-French industry code Industry name Number of LBO targets

Number of firm years

Proportion of LBO targets to firm years (%)

1 Agriculture 2 336 0.60% 2 Food Products 10 1,292 0.77% 3 Candy & Soda 1 237 0.42% 4 Beer & Liquor 2 307 0.65% 6 Recreation 9 843 1.07% 7 Entertainment 10 1,342 0.75% 8 Printing and Publishing 10 1,023 0.98% 9 Consumer Goods 16 1,484 1.08% 10 Apparel 5 1,134 0.44% 11 Healthcare 28 2,361 1.19% 12 Medical Equipment 10 3,228 0.31% 13 Pharmaceutical Products 8 4,786 0.17% 14 Chemicals 7 1,624 0.43% 15 Rubber and Plastic Products 7 684 1.02% 16 Textiles 5 450 1.11% 17 Construction Materials 15 1,821 0.82% 18 Construction 7 1,073 0.65% 19 Steel Works Etc 6 1,156 0.52% 21 Machinery 16 3,024 0.53% 22 Electrical Equipment 15 2,845 0.53% 23 Automobiles and Trucks 12 1,102 1.09% 24 Aircraft 1 423 0.24% 28 Mining 1 425 0.24% 30 Petroleum and Natural Gas 10 3,726 0.27% 32 Communication 18 2,687 0.67% 33 Personal Services 22 1,040 2.12% 34 Business Services 120 12,708 0.94% 35 Computers 14 3,092 0.45% 36 Electronic Equipment 21 5,133 0.41% 37 Measuring and Control Equipment 8 1,863 0.43% 38 Business Supplies 4 909 0.44% 39 Shipping Containers 4 280 1.43% 40 Transportation 23 2,060 1.12% 41 Wholesale 32 4,182 0.77% 42 Retail 57 4,886 1.17% 43 Restaraunts, Hotels, Motels 42 2,264 1.86% 48 Other 8 1,420 0.56% Total 586 79,450 0.74%

Page 34: What does an LBO signal for the target’s industry? · PDF fileWhat does an LBO signal for the target’s industry? Abstract This study examines the information that an LBO signals

32

Panel C: Characteristics of LBOs and acquiring PE funds/firms Variable N mean sd p5 p25 p50 p75 p95 Characteristics of LBO deal Transaction value ($B) 543 1.138 3.174 0.011 0.066 0.221 0.802 4.987 Bid premium 524 0.375 0.730 -0.014 0.153 0.275 0.423 0.903 Target 3-day announcement CARs 583 0.206 0.248 -0.042 0.060 0.168 0.290 0.586 Lead M&A wave 578 0.253 0.435 0.000 0.000 0.000 1.000 1.000 Industry peers of LBO target

Number of industry peers 586 33.3 67.8 0.0 3.0 9.0 35.0 136.0 Probability industry peers acquired

in one year 537 5.26% 8.13% 0.00% 0.00% 2.25% 8.51% 18.75% Probability industry peers acquired

in three years 537 12.77% 14.22% 0.00% 0.00% 11.11% 19.51% 34.29% Characteristics of PE funds

Number of investors 586 1.408 0.917 1.000 1.000 1.000 2.000 3.000 Number of PE investors 586 0.995 0.955 0.000 0.000 1.000 1.000 2.000 Non-PE deal 586 0.092 0.289 0.000 0.000 0.000 0.000 1.000 MBO deal 586 0.282 0.450 0.000 0.000 0.000 1.000 1.000 Age of lead PE firm 404 15.512 10.417 1.000 8.000 15.000 21.000 34.000 Capital under management of lead PE firm ($B) 395 21.099 30.683 0.149 1.867 7.171 26.157 96.032 Sequence order of lead PE fund 388 7.892 10.105 1.000 2.000 4.000 9.000 30.000 Size of lead PE fund ($B) 388 4.030 5.837 0.088 0.535 1.577 4.825 16.611 Number of LBOs lead PE firm participated 402 1.485 2.343 0.000 0.000 1.000 2.000 6.000 during past 3 years

Aggregate size of LBO deals lead PE firm 402 6.541 20.268 0.000 0.000 0.062 1.510 45.519 participated during past 3 years ($B)

Page 35: What does an LBO signal for the target’s industry? · PDF fileWhat does an LBO signal for the target’s industry? Abstract This study examines the information that an LBO signals

33

Table 2: Acquisitions of industry peers of LBO targets This table investigates the likelihood that a firm becomes an acquisition target in a year. It presents the log-odds of the explanatory variables on the likelihood of becoming a target. The sample includes firm-years from 1991 to 2012 covered by the Compustat databases excluding financial firms and utilities. All model specifications employ standard errors robust to 4-digit-SIC industry-level clustering. The associated t-statistics are reported in the parentheses below each coefficient. Superscripts ***, **, and * correspond to statistical significance at the one, five, and ten percent levels, respectively. Tobin’s Q equals assets (Compustat item at) – common equity (ceq) + stock price (prcc_f) times common shares outstanding (csho) – deferred taxes (txdb), divided by total assets (at). Asset structure is the ratio of property, plant, and equipment (ppe) to total assets (at). Cash is the ratio of cash and short-term investments (che) to total assets (at). Market capitalization equals stock price (prcc_f) times common shares outstanding (csho). Book leverage equals debt in current liabilities (dlc) + long-term liabilities (dltt), divided by total assets (at). ROA is the ratio of income before extraordinary items (ib) to total assets (at). All accounting variables are measured at the end of fiscal year t, and are adjusted with respect to the 2-digit-SIC industry median. Blockholder ownership is the ownership of institutional investors with ownership above five percent. Deregulation takes the value of one if there is a deregulation to the industry in year t, and zero otherwise. Economic Shock is the first principal component of the absolute value of net income/sales(t-1), sale/assets(t-1), R&D/assets(t-1), CAPX/assets(t-1), annual employee growth, OIBDP/assets(t-1), and annual sales growth. (1) (2) (3) (4) (5) (6)

Being

Being LBO

Being non-LBO

Dep. var. Target in year t+1

target in year t+1

target in year t+1 Industry peer targeted in 0.3013***

0.3455**

0.2905**

LBO in year t (2.770)

(2.514)

(2.328) Industry peer targeted in

0.2973***

0.0254

0.3259***

non-LBO M&A in year t

(5.015)

(0.219)

(5.124) Tobin's Q -0.1250*** -0.1268***

-0.3669*** -0.3657***

-0.1121*** -0.1143***

(-7.775) (-7.854)

(-5.042) (-4.948)

(-7.090) (-7.215)

Asset structure -0.0091 0.0000

-0.0133 -0.0306

-0.0140 -0.0024

(-0.096) (0.000)

(-0.047) (-0.108)

(-0.143) (-0.024)

Log cash 0.0256** 0.0229*

-0.0112 -0.0082

0.0324*** 0.0292**

(2.176) (1.936)

(-0.331) (-0.242)

(2.624) (2.362)

Log market capitalization -0.0354*** -0.0356***

-0.0570** -0.0582**

-0.0323*** -0.0325***

(-3.641) (-3.642)

(-2.065) (-2.097)

(-3.177) (-3.189)

Book leverage 0.3041*** 0.3007***

0.8753*** 0.8808***

0.2415*** 0.2377***

(3.982) (3.980)

(3.371) (3.400)

(3.133) (3.123)

ROA -0.1352** -0.1035*

1.0258*** 1.0014***

-0.1936*** -0.1581***

(-2.418) (-1.830)

(3.508) (3.431)

(-3.347) (-2.740)

Blockholder ownership 0.3454*** 0.3534***

0.7795*** 0.7762***

0.3038*** 0.3126***

(7.993) (8.226)

(6.389) (6.358)

(6.550) (6.772)

Deregulation 0.2167 0.2244

0.1558 0.1781

0.2187 0.2259

(1.238) (1.412)

(0.262) (0.304)

(1.292) (1.467)

Economic Shock 0.1125* 0.0236 -0.2462 -0.2644 0.1578** 0.0603 (1.697) (0.277) (-0.842) (-0.899) (2.522) (0.792) Constant -3.0601*** -3.1773***

-5.2598*** -5.2219***

-3.2231*** -3.3583***

(-34.543) (-33.340)

(-23.346) (-22.777)

(-35.006) (-33.636)

Year fixed effects Yes Yes

Yes Yes

Yes Yes Observations 80,564 80,564

80,564 80,564

80,564 80,564

Pseudo R2 0.0182 0.0193 0.0632 0.0619 0.0161 0.0177

Page 36: What does an LBO signal for the target’s industry? · PDF fileWhat does an LBO signal for the target’s industry? Abstract This study examines the information that an LBO signals

34

Table 3: Aggregate relationship between merger activity and leveraged buyouts This table investigates the relationship between aggregate merger activity and aggregate leveraged buyout activity. It presents the results of vector autoregression analysis between quarterly aggregated non-LBO merger activity and LBO activity using 4 lags. The sample is all non-rumored mergers and acquisitions (form codes, AM, M, and AA) covered by Thompson-Reuters SDC database from 1991 through 2012, excluding financials and utilities. The associated t-statistics are reported in the parentheses below each coefficient. Superscripts ***, **, and * correspond to statistical significance at the one, five, and ten percent levels, respectively. M&A activity in quarter t is equal to the natural log of 1 plus the number of non-rumored, non-LBO mergers announced in quarter t. LBO activity in quarter t equals the natural log of 1 plus the number of LBO’s announced in quarter t. Macroeconomic controls are estimated as endogenous in specifications (3)-(4) and include real GDP growth, the GDP price deflator, total industrial production, total durable consumer goods production, total business equipment production, and the 10-year Treasury yield obtained from FRED and measured at the quarterly level. (1) (2) (3) (4)

M&A activity in quarter t

LBO activity in quarter t

M&A activity in quarter t

LBO activity in quarter t

LBO activity in quarter t-1 0.0855** 0.7716*** 0.1137** 0.5501***

(2.0087) (7.2959) (2.3406) (4.6897)

LBO activity in quarter t-2 -0.0598 0.0607 -0.0846 -0.043

(-1.1075) (0.4525) (-1.5602) (-0.3283)

LBO activity in quarter t-3 -0.0435 -0.0074 0.0443 -0.1059

(-0.8072) (-0.0551) (0.8852) (-0.8761)

LBO activity in quarter t-4 0.0128 0.1104 0.0135 0.1565

(0.2904) (1.0086) (0.3322) (1.5993)

M&A activity in quarter t-1 0.8525*** 0.1035 0.7082*** -0.1545

(7.3056) (0.3567) (6.3104) (-0.5702)

M&A activity in quarter t-2 0.1628 0.2926 0.0953 0.2697

(1.1324) (0.8191) (0.7382) (0.8655)

M&A activity in quarter t-3 -0.0779 -0.0369 -0.1044 0.0685

(-0.5894) (-0.1124) (-0.8475) (0.2303)

M&A activity in quarter t-4 -0.0144 -0.3936 0.2136** -0.328

(-0.1396) (-1.5307) (1.9664) (-1.2502)

Constant 0.6054** 0.5503 0.0878 -0.5702

(2.0877) (0.7635) (0.1238) (-0.333)

Macroeconomic Controls No No Yes Yes Observations 88 88 Pseudo R2 0.876 0.781 0.926 0.876

Page 37: What does an LBO signal for the target’s industry? · PDF fileWhat does an LBO signal for the target’s industry? Abstract This study examines the information that an LBO signals

35

Table 4: Are LBOs early or late in an industry merger sequence? This table investigates the location of LBOs within industry merger sequences and the likelihood that an LBO target is the first mover in an industry merger sequence. Summary statistics of sequences (containing more than 1 deal) are reported in Panel A. Specifications (1)-(2) of Panels B and C present OLS regression coefficients of the explanatory variables on Log Merger sequence; defined as the natural log of the running count of mergers in a 4-digit SIC industry where there is not more than 365 days between mergers, else the count restarts at 1 between 1990 and 2012. Specifications (3)-(4) of Panels B and C presents the log-odds of the explanatory variables on the likelihood of being the first mover in a merger sequence as estimated in a logit regression. First mover is an indicator variable equal to 1 if a given deal’s merger sequence defined above is equal to 1 and the total number of deals in the sequence is at least 2 between 1990 and 2012. Specifications (2) and (4) restrict the sample to deals located in a merger sequence with at least 2 deals. The sample includes merger targets from 1991 to 2012 covered by both the CRSP and the Compustat databases, excluding financials and utilities, in Panel B. All public and private merger targets, excluding financials and utilities, are included in Panel C. All model specifications employ standard errors robust to 4-digit-SIC industry-level clustering. The associated t-statistics are reported in the parentheses below each coefficient. Superscripts ***, **, and * correspond to statistical significance at the one, five, and ten percent levels, respectively. LBO is an indicator equal to 1 if the deal is an LBO. Other control variables are as defined in Table 2. Panel A: Summary statistics of merger sequences (>1 Deal) Variable N mean sd p5 p25 p50 p75 p95 Number of deals 666 7.317 24.258 2.0 2.0 3.0 4.0 23.0 Number of years 666 2.461 2.584 1.0 1.0 2.0 2.0 7.0 LBO sequence 666 0.309 0.463 0.0 0.0 0.0 1.0 1.0

Page 38: What does an LBO signal for the target’s industry? · PDF fileWhat does an LBO signal for the target’s industry? Abstract This study examines the information that an LBO signals

36

Panel B: Public target firms (1) (2) (3) (4) Dep. var. Log merger sequence First mover LBO -0.4209*** -0.4415*** 0.4617*** 0.5025***

(-3.690) (-3.907) (3.452) (3.458)

Tobin's Q 0.1095*** 0.0864*** -0.1054** -0.1155***

(3.460) (2.941) (-2.549) (-2.762)

Asset structure 0.2761 0.3019 -0.5975* -0.7042*

(0.959) (1.014) (-1.787) (-1.922)

Log cash 0.1267*** 0.1089** -0.0630* -0.0984**

(2.628) (2.366) (-1.706) (-2.406)

Log market capitalization 0.0172 -0.0023 0.0155 -0.0001

(0.757) (-0.101) (0.504) (-0.003)

Book leverage -0.0174 -0.0420 -0.0054 -0.0041

(-0.131) (-0.322) (-0.025) (-0.017)

ROA -0.7312*** -0.4226*** 0.1848 0.3993**

(-7.198) (-4.359) (0.948) (2.022)

Blockholder ownership -0.3410** -0.3806** 0.1982 0.2420

(-2.023) (-2.116) (0.578) (0.694)

Deregulation 0.3081 0.2938 -0.6032* -0.7217* (1.213) (1.246) (-1.707) (-1.950) Economic shock 0.7327 0.6132 -0.2843 -0.3970 (1.462) (1.250) (-1.021) (-1.239) Constant 0.0956 0.4527* -1.1455*** -0.5067**

(0.380) (1.744) (-6.366) (-2.362)

Year fixed effects Yes Yes Yes Yes Observations 5,436 4,452 5,436 4,452 Adjusted and Pseudo R2 0.147 0.176 0.031 0.045 Panel C: All target firms (1) (2) (3) (4) Dep. var. Log merger sequence First mover LBO -0.6964*** -0.6632*** 0.4423*** 0.4519***

(-6.091) (-5.904) (4.263) (4.348)

Constant 3.3125*** 2.9672*** -4.5367*** -4.5194***

(23.386) (19.700) (-23.950) (-23.615)

Year fixed effects Yes Yes Yes Yes Observations 136,253 134,959 136,253 134,959 Adjusted and Pseudo R2 0.162 0.171 0.014 0.014

Page 39: What does an LBO signal for the target’s industry? · PDF fileWhat does an LBO signal for the target’s industry? Abstract This study examines the information that an LBO signals

37

Table 5: Are Acquisitions following LBOs different? Difference-in-difference analysis This table reports the difference-in-difference analysis of merger sequences containing LBO targets and sequences not containing LBO targets, comparing the characteristics of non-LBO targets pre-LBO (or pre-pseduo-LBO) and post-LBO. Panel A reports a comparison of the average characteristics of targets within sequences containing an LBO target, where deals are classified as Follows LBO in Wave as non-LBO targets following the first LBO targeted within a given merger sequence (as defined in Table 4) and as Precedes LBO in Sequence as non-LBO targets preceding the first LBO targeted within a given merger sequence. Panel B reports the average means, counts, differences, and t-statistics over 500 simulations for merger sequences that do not contain an LBO target, but a pseudo-LBO target is created. The pseudo-LBO target in each sequence not containing an LBO target is chosen randomly using a normal distribution with the mean sequence location and the standard deviation of the sequence location obtained from the waves containing an LBO. Two groups (preceding the pseudo-LBO and following the pseudo-LBO) are then formed and compared similar to Panel A. This process is repeated 500 times, with the averages presented below. Panel C presents the differences of deals following or preceding the LBO (or pseudo-LBO) by taking the difference of means between Panel A and Panel B. Deal completion is an indicator variable equal to one if the SDC indicated the deal was completed. Bidder (Target) CAR is the three-day cumulative abnormal returns to the bidder (target) around the acquisition announcement. Abnormal returns are computed based on the market model (Brown and Warner, 1980), with beta estimated over the period from 42 to 252 days prior to the announcement date. Superscripts ***, **, and * correspond to statistical significance at the one, five, and ten percent levels, respectively. Panel A: Waves containing LBO targets

Follows LBO in Wave Precedes LBO in Wave

Mean N Mean N Difference t-statistic Deal completion 0.7856 1866 0.6763 1282 0.1094*** 6.932 Target CAR 0.2242 1850 0.1829 1274 0.0413*** 3.627 Bid premium 0.4661 1592 0.4093 1044 0.0568** 2.182 Bidder CAR -0.0246 1069 -0.0097 472 -0.015** -2.547 Panel B: Waves not containing LBO targets

Follows Pseudo-LBO in Wave

Precedes Pseudo-LBO in Wave

Mean N Mean N Difference t-statistic Deal completion 0.7000 619.4 0.6734 644.1 0.0267 1.021 Target CAR 0.1913 611.9 0.1902 638.8 0.0012 0.073 Bid premium 0.4389 480.6 0.4387 491.4 0.0002 0.004 Bidder CAR -0.0185 316.8 -0.0067 362.7 -0.0118 -1.400 Panel C: Difference (Waves containing LBO targets - Waves not containing LBO targets)

Follows LBO in Wave Precedes LBO in Wave

Difference t-statistic Difference t-statistic

Difference- in-

Difference t-statistic Deal completion 0.0856*** 4.363 0.0029 0.127 0.0827*** 2.694 Target CAR 0.0329** 2.166 -0.0073 -0.551 0.0402** 2.092 Bid premium 0.0271 0.823 -0.0295 -0.903 0.0566 1.357 Bidder CAR -0.0061 -0.882 -0.0030 -0.380 -0.0031 -0.343

Page 40: What does an LBO signal for the target’s industry? · PDF fileWhat does an LBO signal for the target’s industry? Abstract This study examines the information that an LBO signals

38

Table 6: Stock returns to industry peers of LBO targets around LBO announcement This table presents abnormal stock returns to industry peers of LBO targets. Abnormal returns are computed based on the market model (Brown and Warner, 1980), with beta estimated over the period from 42 to 252 days prior to the announcement date. We require at least twenty data points during the estimation window to have an accurate beta estimation. Our sample includes LBOs targeting public US firms announced between 1991 and 2012. Day N Abnormal return (%) T-statistic -5 537 -0.17** (-2.124) -4 537 -0.03 (-0.391) -3 537 -0.07 (-0.844) -2 537 -0.07 (-0.846) -1 537 0.20 (1.545) 0 537 0.21** (2.292) 1 537 -0.01 (-0.119) 2 537 0.08 (0.804) 3 537 0.07 (0.904) 4 537 0.09 (1.103) 5 537 0.03 (0.353) (-1, 1) 537 0.39** (2.333) (1, 21) 537 0.68* (1.673)

Page 41: What does an LBO signal for the target’s industry? · PDF fileWhat does an LBO signal for the target’s industry? Abstract This study examines the information that an LBO signals

39

Table 7: Long-run abnormal returns to industry peers of LBO targets At the beginning of each month from January 1991 to December 2012, we form a portfolio of stocks whose 4-digit SIC industry peer was a target in our sample LBO deals announced during the preceding 12 months. The industry peers are excluded from the monthly portfolio after the month in which the LBO deal is withdrawn. Panel A summarizes the monthly portfolio returns. Panel B presents the OLS regression results where the dependent variable is the equal-weighted (EW) or value-weighted (VW) portfolio return in excess of the risk-free interest rate. The independent variables are the four risk factors—MKT, SMB, HML, and UMD—constructed by Fama and French (1993) and Carhart (1997) plus the “takeover factor” constructed by Cremers, Nair, and John (2009). The first two rows of Panel C present estimated Jensen’s alpha based on the Fama-French-Carhart four factors over two sample periods, 1991-2001 and 2002-2012. We also divide the portfolio stocks into two portfolios based on market capitalization and present the estimated alpha in the last two rows of Panel C. The portfolio of small (large) stocks contains those with market capitalization below (above) the median of all NYSE-listed stocks. All model specifications employ robust standard errors. The associated t-statistics are reported in the parentheses below each coefficient. Superscripts ***, **, and * correspond to statistical significance at the one, five, and ten percent levels, respectively. Panel A: Portfolio returns N mean sd p5 p25 p50 p75 p95 Number of stocks in portfolio 262 452.7 323.5 64.0 222.0 393.5 580.0 1144.0 Equal-weighted portfolio return 262 1.76% 8.09% -10.83% -2.79% 2.21% 6.00% 13.12% Value-weighted portfolio return 262 1.35% 6.50% -9.84% -2.34% 1.34% 5.03% 11.28% Panel B: Jensen’s Alpha (1) (2) (3) (4) Dep. var. Excess portfolio return EW VW EW VW MKT 1.05*** 1.16*** 1.04*** 1.17***

(18.22) (28.27) (18.89) (28.58)

SMB 1.07*** 0.09 1.03*** 0.09

(12.62) (1.35) (12.31) (1.37)

HML -0.38*** -0.57*** -0.51*** -0.56***

(-4.31) (-8.45) (-5.01) (-7.16)

UMD -0.35*** 0.05 -0.28*** 0.04

(-3.98) (1.19) (-3.71) (0.94)

Takeover factor

0.38*** -0.02

(3.51) (-0.24)

Alpha 0.99*** 0.58*** 0.65*** 0.60***

(4.24) (3.55) (3.08) (3.43)

Observations 262 262 262 262 Adjusted R2 0.840 0.838 0.850 0.838

Page 42: What does an LBO signal for the target’s industry? · PDF fileWhat does an LBO signal for the target’s industry? Abstract This study examines the information that an LBO signals

40

Panel C: Estimated alpha by time period and by firm size (1) (2) (3) (4) EW VW

EW VW

Robustness over time periods

1991 - 2001

2002 - 2012 Alpha 1.62*** 0.94***

0.48** 0.18

(3.50) (3.39)

(2.40) (1.05)

Robustness by firm size

Small

Large Alpha 0.71*** 0.62***

0.60*** 0.60***

(3.57) (3.23) (3.29) (3.28)

Page 43: What does an LBO signal for the target’s industry? · PDF fileWhat does an LBO signal for the target’s industry? Abstract This study examines the information that an LBO signals

41

Table 8: Long-run returns to industry peers of non-LBO merger and acquisition targets At the beginning of each month from January 1980 to December 2012, we form a portfolio of stocks whose listed 4-digit SIC industry peer was the target of non-LBO mergers and acquisitions announced during the preceding 12 months. The industry peers are excluded from the monthly portfolio after the month in which the merger and acquisition is withdrawn. This table presents the estimated Jensen’s alpha for the portfolios. The alpha is estimated using the OLS regression where the dependent variable is the equal- or value-weighted portfolio return in excess of the risk-free interest rate and the independent variables are the four risk factors—MKT, SMB, HML, and UMD—constructed by Fama and French (1993) and Carhart (1997). All model specifications employ robust standard errors. The associated t-statistics are reported in the parentheses below each coefficient. Superscripts ***, **, and * correspond to statistical significance at the one, five, and ten percent levels, respectively. (1) (2) EW VW Dep. var. Excess portfolio return MKT 0.99*** 0.99***

(28.17) (57.23)

SMB 1.07*** 0.13***

(16.41) (4.21)

HML -0.21*** -0.29***

(-3.11) (-9.73)

UMD -0.28*** -0.01

(-4.83) (-0.65)

Alpha 0.31** 0.09

(2.08) (1.36)

Observations 394 394 Adjusted R2 0.885 0.936

Page 44: What does an LBO signal for the target’s industry? · PDF fileWhat does an LBO signal for the target’s industry? Abstract This study examines the information that an LBO signals

42

Table 9: Announcement CARs and long-run abnormal returns to industry peers of LBO targets: By the probability that industry peers of LBO target are acquired after the LBO Panel A presents average 3-day CARs to industry peers of the LBO target around the LBO announcement, grouped by whether a variable of interest is above or below the median. The variables of interest include the probability that rivals of LBO target are acquired in the one or three years after LBO and whether the LBO leads a merger wave. An LBO leads a merger wave if it is announced over the 24 months prior to or over the first 12 months after the beginning of a merger wave in the same industry as defined by the methods in Harford (2005). At the beginning of each month from January 1991 to December 2012, we form a portfolio of stocks whose 4-digit SIC industry peer was a target in our sample LBO deals announced during the preceding 12 months. The industry peers are excluded from the monthly portfolio after the month in which the LBO deal is withdrawn. We further divide the portfolio into two based on whether a variable of interest is above or below the median. Loan spread is the quarterly commercial loan rate in excess of the federal funds rate. Panel B presents the estimated Jensen’s alpha for the portfolios. The alpha is estimated using the OLS regression where the dependent variable is the equal- or value-weighted portfolio return in excess of the risk-free interest rate and the independent variables are the four risk factors—MKT, SMB, HML, and UMD—constructed by Fama and French (1993) and Carhart (1997). All model specifications employ robust standard errors. The associated t-statistics are reported in the parentheses below each coefficient. Superscripts ***, **, and * correspond to statistical significance at the one, five, and ten percent levels, respectively. Panel A: 3-day CARs to industry peers of the LBO target (1) (2) (3) Low High Difference (High - Low) Probability industry peers acquired in 1 year 0.74** 0.04 -0.70**

(2.51) (0.26) (-2.09)

Probability industry peers acquired in 3 years 0.72** 0.04 -0.68**

(2.54) (0.23) (-2.08)

Lead merger wave 0.55** -0.09 -0.64** (2.58) (-0.36) (-1.98) Panel B: Estimated Jensen’s alpha (1) (2) (3) (4) (5) (6)

Low High Difference

EW VW EW VW EW VW Probability industry peers acquired in 1 year 0.59* 0.35

0.92*** 0.57***

0.33 0.23

(1.96) (1.37)

(3.71) (2.84)

(0.85) (0.70)

Probability industry peers acquired in 3 years 0.57** 0.36

0.96*** 0.63***

0.39 0.27

(1.97) (1.51)

(3.71) (3.06)

(1.01) (0.85)

Lead merger wave 0.57* 0.04

1.12*** 0.83**

0.54 0.80*

(1.97) (0.14)

(3.16) (2.33)

(1.19) (1.80)

Loan spread 1.33*** 0.71***

0.63*** 0.30

-0.70 -0.41 (2.96) (2.66) (2.75) (1.46) (-1.39) (-1.24)

Page 45: What does an LBO signal for the target’s industry? · PDF fileWhat does an LBO signal for the target’s industry? Abstract This study examines the information that an LBO signals

43

Table 10: Changes in leverage, strategic decisions, and governance of peers of an LBO target Panel A presents OLS regression results where the dependent variables are the firm’s cash holding, book leverage, and market leverage in year t+1. Market leverage equals debt in current liabilities (Compustat item dlc) + long-term liabilities (dltt), divided by market value of assets: assets (at) – common equity (ceq) + stock price (prcc_f) times common shares outstanding (csho). Panel B presents OLS and logit regression results where the dependent variables are annual change in R&D, and alliance, strategic alliance, and spin off indicators. The data on alliance and spin off are retrieved from the SDC database. Panel C presents OLS regression results where the dependent variables are changes in five proxies for corporate governance. Board size is the number of board directors. Board independent is the fraction of independent directors. The E/G index is based on the number of antitakeover provisions as constructed by Bebchuk, Cohen, and Ferrell (2009) and Gompers, Ishii, and Metrick (2002). CEO pay is the total direct compensation to the CEO as reported in the ExecuComp database. Panel D presents OLS regression results where the dependent variable is a firm’s share repurchases in year t+1. Following Banyi, Dyl, and Kahle (2008), share repurchase is defined as purchase of common and preferred stock (Compustat item prstkc) minus changes in preferred stock (pstk), divided by the firm’s market capitalization at the beginning of the year (csho times prcc_f). See Table 2 for definitions of other variables. All model specifications employ robust standard errors. The associated t-statistics are reported in the parentheses below each coefficient. Superscripts ***, **, and * correspond to statistical significance at the one, five, and ten percent levels, respectively. Panel A: Leverage (1) (2) (3)

Log cash Book leverage Market leverage

Dep. var. t+1 t+1 t+1 Industry peer targeted in 0.0046 0.0012 -0.0013 LBO in past 3 years (0.389) (1.018) (-0.859) Tobin's Q 0.0253*** -0.0010*** -0.0107***

(6.526) (-2.640) (-8.371)

Asset structure -0.2505*** 0.0154*** 0.0279***

(-7.866) (4.037) (4.776)

Log cash 0.7355*** -0.0044*** -0.0099***

(144.782) (-9.360) (-12.650)

Log market capitalization -0.0067** 0.0016*** -0.0074***

(-2.126) (4.934) (-13.129)

Book leverage -0.4188*** 0.8464*** 0.5425***

(-14.697) (151.948) (44.301)

ROA -0.1408*** -0.0195*** 0.0169***

(-10.222) (-7.895) (7.863)

Blockholder 0.0440*** -0.0041*** 0.0049***

(5.010) (-4.047) (4.134)

Deregulation -0.0597* -0.0185*** -0.0150***

(-1.681) (-3.550) (-3.386)

Economic shock 0.0088 0.0035* 0.0017

(0.460) (1.666) (0.679)

Constant -0.0622*** 0.0063* 0.0170***

(-3.899) (1.737) (8.232)

Year fixed effects Yes Yes Yes Observations 72,772 72,919 72,835 Adjusted R2 0.620 0.703 0.591

Page 46: What does an LBO signal for the target’s industry? · PDF fileWhat does an LBO signal for the target’s industry? Abstract This study examines the information that an LBO signals

44

Panel B: Strategic decisions (1) (2) (3) (4)

Change in Strategic Alliance Spinoff

Dep. var. R&D, t+1 alliance, t+1 t+1 t+1 Industry peer targeted in 0.0103* 0.5005*** 0.3741** -0.0308 LBO in past 3 years (1.870) (2.594) (2.187) (-0.181) Tobin's Q 0.0040* 0.0124 -0.0078 -0.1777***

(1.828) (0.886) (-0.550) (-3.773)

Asset structure -0.0154** -0.5225*** -0.4020*** -0.5825

(-2.042) (-4.180) (-3.935) (-1.469)

Log cash -0.0021 0.0917*** 0.0654*** 0.0098

(-1.532) (4.776) (3.999) (0.230)

Log market capitalization -0.0082*** 0.4679*** 0.4892*** 0.5541***

(-3.580) (26.400) (31.054) (18.307)

Book leverage 0.0043 0.1780* 0.1824** 0.1041

(0.282) (1.948) (2.288) (0.382)

ROA 0.2414*** -0.9056*** -0.9090*** -0.7425***

(5.446) (-8.191) (-8.555) (-3.562)

Blockholder -0.0116** -0.1566*** -0.1601*** -0.0443

(-2.300) (-3.959) (-4.511) (-0.445)

Deregulation -0.0029 -0.2850 -0.0957 0.4261

(-0.428) (-1.286) (-0.523) (1.326)

Economic shock -0.0189* -0.1362 -0.2028 -0.0866

(-1.817) (-0.518) (-0.903) (-0.423)

Constant 0.0140 -2.4080*** -1.9480*** -5.5047***

(1.270) (-19.726) (-19.690) (-21.535)

Year fixed effects Yes Yes Yes Yes Observations 73,141 80,564 80,564 70,693 Adjusted/Pseudo R2 0.012 0.129 0.130 0.0959

Page 47: What does an LBO signal for the target’s industry? · PDF fileWhat does an LBO signal for the target’s industry? Abstract This study examines the information that an LBO signals

45

Panel C: Corporate governance (1) (2) (3) (4) (5)

Change in Change in Change in Change in Change in

board size independence E index G index CEO pay

Dep. var. t+1 t+1 t+1 t+1 t+1 Industry peer targeted in 0.0049** -0.0024* 0.0042*** 0.0149** -0.0031 LBO in past 3 years (2.065) (-1.715) (3.260) (2.344) (-0.365) Tobin's Q 0.0033*** 0.0004 0.0012*** 0.0013 0.0103***

(5.038) (0.945) (2.752) (0.970) (2.586)

Asset structure -0.0013 -0.0017 0.0003 -0.0090 -0.0128

(-0.280) (-0.468) (0.091) (-0.918) (-0.657)

Log cash 0.0012 -0.0006 0.0004 0.0020 -0.0063**

(1.581) (-1.139) (1.066) (1.540) (-2.008)

Log market capitalization -0.0007 0.0001 -0.0017*** -0.0037*** -0.0028

(-1.327) (0.284) (-5.498) (-3.462) (-1.212)

Book leverage 0.0146** 0.0019 0.0017 0.0129 -0.0375

(2.441) (0.513) (0.741) (1.523) (-1.647)

ROA 0.0209** 0.0039 -0.0027 -0.0014 0.0855***

(2.423) (0.679) (-0.897) (-0.130) (2.586)

Blockholder -0.0010 -0.0009 -0.0012 0.0060 0.0198*

(-0.347) (-0.430) (-1.013) (1.380) (1.671)

Deregulation 0.0206 -0.0190* -0.0025 0.0091 0.0218

(1.388) (-1.674) (-0.417) (0.691) (0.394)

Economic shock 0.0069** 0.0017 0.0008 0.0069 -0.0009

(2.451) (0.783) (0.573) (1.311) (-0.081)

Constant -0.0054 0.0111** 0.0032** 0.0315 0.0491

(-1.104) (2.581) (2.229) (1.192) (1.074)

Year fixed effects Yes Yes Yes Yes Yes Observations 14,505 14,505 19,036 18,065 21,650 Adjusted R2 0.005 0.015 0.068 0.172 0.010

Page 48: What does an LBO signal for the target’s industry? · PDF fileWhat does an LBO signal for the target’s industry? Abstract This study examines the information that an LBO signals

46

Panel D: Share repurchases (1)

Share repurchase

Dep. var. t+1 Industry peer targeted in 0.0011* LBO in past 3 years (1.872) Tobin's Q -0.0019***

(-11.797)

Asset structure -0.0058***

(-3.873)

Log cash 0.0011***

(5.870)

Log market capitalization 0.0019***

(11.372)

Book leverage 0.0019

(1.367)

ROA 0.0016*

(1.903)

Blockholder 0.0019***

(4.826)

Deregulation 0.0030**

(2.061)

Economic shock -0.0029*

(-1.858)

Constant 0.0183***

(20.064)

Year fixed effects Yes Observations 68,439 Pseudo R2 0.032

Page 49: What does an LBO signal for the target’s industry? · PDF fileWhat does an LBO signal for the target’s industry? Abstract This study examines the information that an LBO signals

47

Table 11: Additional difference-in-difference analysis This table reports the difference-in-difference analysis of merger waves containing LBO targets and waves not containing LBO targets, comparing the characteristics of non-LBO targets pre-LBO (or pre-pseduo-LBO) and post-LBO. Panel A reports a comparison of the average characteristics of targets within waves containing an LBO target, where deals are classified as Follows LBO in Wave as non-LBO targets following the first LBO targeted within a given merger sequence (as defined in Table 4) and as Precedes LBO in Wave as non-LBO targets preceding the first LBO targeted within a given merger sequence. Panel B reports the average means, counts, differences, and t-statistics over 500 simulations for merger sequences that do not contain an LBO target, but a pseudo-LBO target is created. The pseudo-LBO target in each wave not containing an LBO target is chosen randomly using a normal distribution with the mean sequence location and the standard deviation of the sequence location obtained from the waves containing an LBO. Two groups (preceding the pseudo-LBO and following the pseudo-LBO) are then formed and compared similar to Panel A. This process is repeated 500 times, with the averages presented below. Panel C presents the differences of deals following or preceding the LBO (or pseudo-LBO) by taking the difference of means between Panel A and Panel B. All cash, All stock, Diversify, and Hostile are indicator variables equal to one if the deal was financed with 100% cash, financed with 100% stock, between to firm not in the same 4-digit SIC industry, and hostile, respectively. Proportion bidder industry’s sales to target industry is the proportion of sales of the bidder industry that are sold to the target industry from the 1992 BEA IO tables. Proportion target industry’s sales to bidder industry is the proportion of sales of the target industry that are sold to the bidder industry from the 1992 BEA IO tables. Bidder industry’s sales to target industry indicator is an indicator variable equal to one if the target industry is responsible for more than one percent of the bidder industry’s sales. Target industry’s sales to bidder industry indicator is an indicator variable equal to one if the bidder industry is responsible for more than one percent of the target industry’s sales. Superscripts ***, **, and * correspond to statistical significance at the one, five, and ten percent levels, respectively. Panel A: Waves containing LBO targets

Follows LBO in Wave Precedes LBO in Wave

Mean N Mean N Difference t-statistic All cash 0.3923 1866 0.4524 1282 -0.0601*** -3.367 All stock 0.2728 1866 0.1732 1282 0.0996*** 6.545 Diversify 0.4357 1866 0.6271 1282 -0.1915*** -10.748 Book leverage 0.0530 1811 0.0667 1259 -0.0137* -1.803 ROA -0.1295 1829 -0.0636 1262 -0.0659*** -6.023 Hostile 0.0123 1866 0.0234 1282 -0.0111** -2.374 Proportion bidder industry’s sales to

target industry 0.0491 1866 0.0306 1283 0.0185*** 9.029 Proportion target industry’s sales to

bidder industry 0.0643 1866 0.0395 1283 0.0248*** 8.673 Bidder industry’s sales to target

industry indicator (=1 if >1%) 0.6233 1866 0.4147 1283 0.2086*** 11.786 Target industry’s sales to bidder

industry indicator (=1 if >1%) 0.6238 1866 0.3632 1283 0.2606*** 14.871

Page 50: What does an LBO signal for the target’s industry? · PDF fileWhat does an LBO signal for the target’s industry? Abstract This study examines the information that an LBO signals

48

Panel B: Waves not containing LBO targets

Follows Pseudo-LBO in Wave

Precedes Pseudo-LBO in Wave

Mean N Mean N Difference t-statistic All cash 0.3220 619.4 0.3157 644.1 0.0063 0.241 All stock 0.2043 619.4 0.2270 644.1 -0.0227 -0.980 Diversify 0.5241 619.4 0.5249 644.1 -0.0008 -0.028 Book leverage 0.0745 604.3 0.0515 631.1 0.0229* 1.843 ROA -0.0770 605.6 -0.0824 631.7 0.0053 0.354 Hostile 0.0378 619.4 0.0403 644.1 -0.0024 -0.225 Proportion bidder industry’s sales to

target industry 0.0329 619.4 0.0319 644.1 0.0010 0.273 Proportion target industry’s sales to

bidder industry 0.0392 611.9 0.0399 638.8 -0.0007 -0.182 Bidder industry’s sales to target

industry indicator (=1 if >1%) 0.4640 480.6 0.4940 491.4 -0.0300 -1.067 Target industry’s sales to bidder

industry indicator (=1 if >1%) 0.4509 316.8 0.4823 362.7 -0.0314 -1.118 Panel C: Difference (Waves containing LBO targets - Waves not containing LBO targets)

Follows LBO in Wave Precedes LBO in Wave

Difference t-statistic Difference t-statistic

Difference- in-

Difference t-statistic All cash 0.0703*** 3.138 0.1368*** 5.813 -0.0665** -2.091 All stock 0.0684*** 3.391 -0.0539** -2.839 0.1223*** 4.456 Diversify -0.0885*** -3.839 0.1022*** 4.327 -0.1907*** -5.736 Book leverage -0.0215** -2.163 0.0152 1.483 -0.0367** -2.512 ROA -0.0525*** -3.539 0.0188 1.522 -0.0713*** -3.881 Hostile -0.0255*** -4.055 -0.0169** -2.069 -0.0086 -0.720 Proportion bidder industry’s sales to

target industry 0.0162*** 5.920 -0.0013 -0.471 0.0175*** 4.254 Proportion target industry’s sales to

bidder industry 0.0251*** 6.802 -0.0004 -0.116 0.0255*** 5.103 Bidder industry’s sales to target

industry indicator (=1 if >1%) 0.1592*** 7.032 -0.0794*** -3.318 0.2386*** 7.177 Target industry’s sales to bidder

industry indicator (=1 if >1%) 0.1729*** 7.642 -0.1191*** -5.058 0.2920*** 8.828

Page 51: What does an LBO signal for the target’s industry? · PDF fileWhat does an LBO signal for the target’s industry? Abstract This study examines the information that an LBO signals

49

Table 12: Selection or causation: All-stock deal likelihood This table investigates whether the likelihood of an announced deal being within-industry and financed using 100% stock and the likelihood that a bidder uses 100% stock to finance a deal. The table presents the log-odds of explanatory variables on the likelihood of a given deal being within-industry and financed through 100% cash in specification (1). Specification (2) presents the log-odds of explanatory variables on the likelihood of a given acquirer finances a bid through 100% cash. The sample in specification (1) includes merger targets from 1991 to 2012 covered by both the CRSP and the Compustat databases, excluding financials and utilities. Specification (2) includes all merger bidders from 1991 to 2012 covered by both the CRSP and the Compustat databases, excluding financials and utilities. Both model specifications employ standard errors robust to 4-digit-SIC industry-level clustering. The associated t-statistics are reported in the parentheses below each coefficient. Superscripts ***, **, and * correspond to statistical significance at the one, five, and ten percent levels, respectively. The explanatory variables are defined in Table 2. (1) (2)

Dep. var.

All-stock, within-industry deal in

year t+1

All-stock bid in

year t+1 Industry peer targeted in LBO in year t 0.4302*** 0.1856*

(3.769) (1.768)

Tobin's Q 0.1456*** 0.1564***

(6.293) (11.233)

Asset structure 0.1451 -0.0686

(0.509) (-0.309)

Log cash 0.1141** 0.1469***

(2.360) (4.832)

Log market capitalization 0.0969*** -0.0405**

(3.746) (-2.354)

Book leverage -0.7568*** -0.7319***

(-2.619) (-5.020)

ROA -0.1998 -0.6035***

(-1.305) (-5.616)

Blockholder ownership -0.4381 -0.5968*

(-1.508) (-1.934)

Deregulation 0.3177 -0.1191 (0.943) (-0.432) Economic shock 0.3904* 0.2319*** (1.939) (2.610) Constant -2.8949*** -2.2243***

(-6.717) (-14.509)

Year fixed effects Yes Yes Observations 4,928 36,269 Pseudo R2 0.079 0.091

Page 52: What does an LBO signal for the target’s industry? · PDF fileWhat does an LBO signal for the target’s industry? Abstract This study examines the information that an LBO signals

50

Table 13: Industry competition and announcement CARs and abnormal returns to industry peers of LBO targets Panel A summarizes the average three-day CARs to industry peers of LBO target firm, grouped by the number of industry peers of the LBO target, or the Herfindahl–Hirschman Index based on sales of all firms in the 4-digit SIC industry of the LBO target at the end of the fiscal year prior to the LBO announcement. Panel B presents OLS regression results where the dependent variables is the average 3-day CARs to industry peers of the LBO target around the LBO announcement. The explanatory variables are the number of industry peers of the LBO target, the Herfindahl–Hirschman Index based on sales of all firms in the 4-digit SIC industry of the LBO target, and the average market-to-book ratio of all firms in the same 4-digit SIC industry at the end of the fiscal year prior to the LBO announcement. All model specifications employ robust standard errors. The associated t-statistics are reported in the parentheses below each coefficient. Superscripts ***, **, and * correspond to statistical significance at the one, five, and ten percent levels, respectively. Our sample includes LBOs targeting public US firms between 1991 and 2012. Panel A: Announcement CARs to industry peers of LBO target (1) (2) (3)

Low High Difference

Number of industry peers 0.94*** -0.07 -1.00***

(2.81) (-0.52) (-2.81)

Herfindahl–Hirschman Index 0.04 0.82** 0.79**

(0.23) (2.56) (2.21)

Panel B: Regression results (1) (2) (3) (4) (5) Dependent variable 3-day CARs of industry peers Number of industry peers, Log -0.37***

-0.32**

(-2.75)

(-2.37) Herfindahl–Hirschman Index, Log

0.47**

0.38*

(2.31)

(1.91)

Market-to-book

-0.14*** -0.07* -0.11**

(-2.86) (-1.78) (-2.44)

Constant 1.31*** 1.05** 0.88*** 1.45*** 1.31***

(2.81) (2.57) (3.14) (3.03) (2.93)

Observations 537 537 537 537 537 R-squared 0.019 0.010 0.009 0.022 0.015

Panel C: Jensen’s Alpha (1) (2) (3) (4) (5) (6)

Low High Difference

EW VW EW VW EW VW Number of industry peers 0.45 -0.20

1.00*** 0.70***

0.55 0.90**

(1.46) (-0.64)

(4.09) (3.62)

(1.41) (2.41)

Herfindahl–Hirschman Index 1.22*** 0.90***

0.57* 0.12

-0.65 -0.78**

(4.48) (4.19)

(1.70) (0.38)

(-1.52) (-2.08)

Page 53: What does an LBO signal for the target’s industry? · PDF fileWhat does an LBO signal for the target’s industry? Abstract This study examines the information that an LBO signals

51

Table 14: Institutional ownership of the LBO target industry around LBOs This table presents OLS regression results where the dependent variables are changes in four variables from four quarters ago: (1) the aggregate ownership of all institutional investors of a 4-digit SIC industry at the end of each quarter from 1991-2012; (2) the aggregate ownership of all hedge funds of a 4-digit SIC industry at the end of each quarter from 2002-2011; (3) the number of unique institutional investors of a 4-digit SIC industry at the end of each quarter from 1991-2012; and (4) the number of unique hedge fund investors of a 4-digit SIC industry at the end of each quarter from 2002-2011. The explanatory dummy variables take the value of one if a firm in the 4-digit SIC industry was targeted in a surrounding quarter, and zero otherwise. All model specifications employ robust standard errors. The associated t-statistics are reported in the parentheses below each coefficient. Superscripts ***, **, and * correspond to statistical significance at the one, five, and ten percent levels, respectively. (1) (2) (3) (4)

Change of Change of

Change of Change of number of number of

institutional hedge fund institutional hedge fund

ownership, ownership, investors, investors,

Dependent variable (Qt-4 to Qt) (Qt-4 to Qt) (Qt-4 to Qt) (Qt-4 to Qt) Quarter -4 before LBO -0.0086** -0.0021 -0.1714 0.4024

(-2.29) (-0.56) (-0.14) (0.57)

Quarter -3 before LBO 0.0027 0.0004 3.4710** 0.0208

(0.83) (0.09) (2.44) (0.02)

Quarter -2 before LBO -0.0010 -0.0024 0.3442 0.0044

(-0.28) (-0.61) (0.26) (0.00)

Quarter -1 before LBO -0.0007 -0.0011 2.4017* -0.3956

(-0.17) (-0.32) (1.69) (-0.54)

Quarter 0 of LBO -0.0098*** 0.0011 2.5216* 2.2444***

(-2.58) (0.27) (1.96) (2.67)

Quarter 1 after LBO -0.0142*** 0.0010 -7.9331*** -3.8195***

(-3.01) (0.29) (-4.09) (-4.39)

Quarter 2 after LBO -0.0012 0.0005 -5.3562*** -4.4281***

(-0.26) (0.07) (-3.08) (-3.96)

Quarter 3 after LBO 0.0008 0.0047 -4.6179*** -2.7809***

(0.17) (1.11) (-3.22) (-3.15)

Quarter 4 after LBO 0.0014 0.0035 -1.4242 -1.5840**

(0.36) (1.18) (-0.89) (-2.01)

Constant -0.0046* -0.0082* 4.2608*** -3.4976***

(-1.79) (-1.71) (8.67) (-5.85)

Industry fixed effects Yes Yes Yes Yes Time fixed effects Yes Yes Yes Yes Observations 82,119 34,086 82,515 34,315 R-squared 0.071 0.018 0.103 0.087